BMI, waist to height ratio and waist circumference as a screening tool for hypertension in hospital out patients: a cross-sectional study

Background: Obesity has become a global epidemic with a rise in noncommunicable diseases. It is now becoming the problem of low- and middle-income countries such as Nepal. Conventional risk factors are present in a high proportion in the Nepalese population. As a routine surveillance or registry system is absent, the actual burden and trend of obesity and hypertension in Nepal are unknown. Hypertension and other cardiovascular diseases can be prevented by detecting risk factors such as obesity and high blood pressure. A simple anthropometric measurement could be used to determine the risk of hypertension. However, the best predictor of hypertension remains contentious and controversial. We aimed to determine the burden of obesity and hypertension and test the ability to determine hypertension through different anthropometric measurements in hospital outpatients in a low-income setting. Methods: This hospital-based cross-sectional descriptive study was conducted from June to December 2019 among 40-69 year outpatients in a tertiary eye and ENT hospital in a semi-urban area of Nepal among a randomly selected sample of 2,256 participants from 6,769 outpatients visited in Health Promotion and risked factor screening service. We performed a correlation analysis to determine the relationship between anthropometric measurements and blood pressure. The area under the receiver operating characteristic (ROC) curve of body mass index (BMI), waist to height ratio (WHtR) and waist circumference (WC) was calculated and compared. Results: The mean (SD) age of the participants was 51.75 (8.47) years. The overall prevalence of obesity and overweight by BMI was 16.09% and 42.20%, respectively. The overall prevalence of abdominal obesity by waist-to-height ratio was 32.76%, which is higher than obesity by BMI. High waist circumference was observed among 66.76% participants, whereas female participants had a very higher prevalence of high waist circumference (77.46%) and male participants (53.73%) (p<0.001). The prevalence of hypertension among the participants with BMI ≥ 25 kg/m2 , WHtR ≥ 0.5 and WC ≥ cutoff values was 45.97%, 42.52% and 45.28%, respectively. The overall prevalence of hypertension

although the best predictor of hypertension remains contentious and controversial. BMI, waist circumference, and waist to height ratio are anthropometric screening tools for predicting hypertension and other cardiovascular diseases [39]. Some studies suggest that waist circumference (WC) and WHtR may be better predictors for hypertension CVD risk [40][41][42][43][44][45], and some studies suggest that BMI and WC are predictors of hypertension [46][47][48]. A meta-analysis suggests that the WC was a better predictor for CVD risks such as hypertension and recommends that it should be used in the clinic and research [49].
Most of the studies on obesity, hypertension and their association were conducted in community settings.
There is a paucity of evidence on the burden of obesity, hypertension, their association and the predictability of hypertension by BMI, WHtR and WC is lacking in the Nepalese setting. Therefore, this study aims to determine the burden of obesity and hypertension and to test the ability to determine hypertension through different anthropometric measurements in hospital outpatients in a low-income setting.

Methods
This hospital-based cross-sectional descriptive study was conducted from June to December 2019 at Hospital for Children Eye, ENT and Rehabilitation Services (CHEERS), Bhaktapur, Nepal. We used systematic random sampling to select the participants. Every third participant aged 40 to 69 visiting the health promotion and risk factor screening service of CHEERS during the study period constituted the study population. The sample size calculation was based on the prevalence of hypertension, 46.7% (P) (steps 2013). The margin of error was 5% (D), the 95% con dence level (Z=1.96) and the 80% response rate. The formula used for sample size calculation is N=(Z^2*P(1-P))/(D^2). The calculated sample size is multiplied by a number of domains to obtain the nal sample size. The number of domains was decided by two age groups and two genders. Based on the calculation, the minimum sample size required was 1,913. All participants were informed about the purpose of the screening service, and informed consent was obtained before anthropometric measurements. Pregnant women and people unable to stand properly were excluded from data analysis for this study.
We followed a standardized protocol at the hospital for anthropometric measurements. Community medicine auxiliaries (CMAs) were trained on an existing protocol for obtaining anthropometric measurements for height, weight, waist circumference and blood pressure. The weight, height and waist circumference were measured on a portable digital weighing scale (Equinox weighing scale), stadiometer (Prestige stadiometer) and constant tension tape, respectively. The participants were asked to remove bulky clothes, shoes and caps before taking measurements. The waist circumference (WC) in cm was measured at the midpoint between the lower edge of the rib cage and the iliac crest. The BMI was then calculated as weight (kg) divided by height squared (m 2 ). The waist-to-height ratio (WHtR) was calculated as WC in cm divided by height in cm. In addition to anthropometric measurements, socio-demographic information, current smoking and drinking habits and history of hypertension (hypertensive medication) were also asked.
The participants were classi ed as hypertensive if their systolic blood pressure (SBP) was ≥140 mmHg and/or diastolic blood pressure (DBP) ≥ 90 mmHg and prehypertensive if systolic blood pressure levels were between 120-139 mmHg and/or diastolic blood pressure levels were between 80-89 mmHg. The participants were also considered hypertensive if they were taking antihypertensive medication, even though their blood pressure measurement was normal. The participants who did not t in all of the above categories were considered normotensive.
Data analysis was performed using R, version 4.0.0. Continuous variables are shown as the mean, standard deviation [3], and categorical variables as frequency and percentage. Independent sample t-tests were performed to compare the mean values of the continuous variables between different groups. We used logistic regression analysis to nd the effect of socio-demographic and different obesity metrics and behavioural risk factors on hypertension separately and in combination. The adjusted odds ratio for hypertensive compared to the nonhypertensive group was analyzed by entering age and sex in a model and risk factors with sociodemographic variables in separate different analysis models. Odds ratios were also reported in 95% con dence intervals. We calculated correlation analysis and calculated Spearman's productmoment correlation coe cient. The area under the receiver operating characteristic (ROC) curve of body mass index (BMI), waist to height ratio (WHtR) and waist circumference (WC) for predicting hypertension and 95% con dence interval (CI) were calculated. The con dence interval, which did not include 0.5, was considered to indicate signi cant results. A p-value <0.05 in all tests was considered signi cant.

Results
This study included 2,256 randomly selected participants from 6,769 people aged 40-69 years who visited health promotion and risk factor screening services in the Hospital for Children Eye ENT and Rehabilitation Services from June to December 2019. The mean age (SD) of the participants was 51.75 (8.47) years.

Prevalence of Obesity and overweight
The mean (SD) BMI was 25.29 (3.81) kg/m 2 and 26.72 (4.44) kg/m 2 among male and female participants, respectively. The mean BMI gradually decreased from younger age to older age groups in both males and females. The overall prevalence of obesity (BMI≥ 30 kg/m2 ) and overweight was 16.09% and 42.20%, respectively. However, female participants had a higher prevalence of obesity (21.4%) than male participants (9.6%) (p-value<0.001). The burden of obesity was higher among the younger age group in both genders.
The overall prevalence of overweight among males and females was 42.8% and 41.7%, respectively, which was not statistically signi cant (p-value = 0.6121). Younger age groups had a signi cantly higher prevalence (p-value <0.001) of overweight in both genders. Table 2 summarizes obesity and overweight according to gender and age groups. The odds ratio for being obese compared with females to males was 2.58 (95% CI: 2.01-3.31), and that of being overweight was 0.95 (95% CI: 0.0.81-1.13) The overall prevalence of abdominal obesity by waist-to-height ratio was 32.76%, which is higher than obesity by BMI. Female participants had a higher prevalence (40.1%) than male participants (23.8%), and the difference was statistically signi cant (p-value < 0.001). The 40-54 age group had a signi cantly higher prevalence of abdominal obesity among female participants (37.59% vs 44.52%, p=0.0195) but not among male participants (23.96% vs 23.66%, p=0.974).     * Correlation is signi cant at the 0.05 level (2-tailed).
Receiver operating characteristic (ROC) analyses were used to determine the relative ability of the three obesity metrics to predict HTN, as depicted in Fig. 1 and Table 5. The areas under the curve (AUCs) were signi cantly higher than 0.5 for BMI (0.570, 95% CI: 0.548-0.592), WC (0.585, 95% CI: 0.563-0.607) and WHtR (0.586, 95% CI: 0.564-0.608). In both genders, the area under the curve was signi cantly higher than 0.5 (P<0.01). In all age groups, the area under the curve was also signi cantly higher than 0.5. Table 5.

Discussion
This study is a result of a health promotion initiative and an opportunistic screening of obesity and hypertension in a tertiary level Eye and ENT hospital in Bhaktapur, Nepal.
In terms of BMI, the overall prevalence of obesity was 16.09% in our study, which is higher than the 2019 STEPS survey Nepal, where 10.4% of the same age group were obese. More than two in ten (21.4%) women were obese, whereas nearly one in ten (9.6%) men were obese in our study, which is nearly twice for both genders (women: 9.5% and men: 5.1%) compared to the Nepal DHS survey of 2016 [24]. In addition, overweight and obesity decreased with increasing age group in our study for both genders. In terms of waist circumference, two-thirds (66.76%) people had their waist circumference above the cutoff point. It is more than twice for the same age group compared to the 2019 STEPS survey. More than three-fourths (77.46%) and more than half (53.73%) of women and men, respectively, had waist circumference more than recommended optimally in our study. The overall prevalence of abdominal obesity by WtHR was 32.76%, and women had a higher prevalence (40.1%) than men (23.8%) in our study. We did not nd any published materials regarding the WtHR study in the Nepali population to compare. However, a study in neighboring India among 20-to 60-year-olds in the capital city showed a high WHtR (>0.50) in 69.9% of the population (69.1% males and 71% females) [53].
All three metrics of obesity in our study showed that women were comparatively obese in a higher proportion than men, which is also supported by other national and international literature [17,23,26]. It is known that abdominal obesity may increase substantially with each pregnancy independent of total body fat [54], which may explain the higher overweight and obesity among women. The higher prevalence of obesity in the present study may be due to study design, a selection bias as people reporting to hospitals may have some or other conditions that can have obesity in the background of their illness. To the best of our knowledge, there is no OPD-based prevalence of obesity data available in Nepal for comparison.
The prevalence of HTN in our study was 40.67%, and men had a slightly higher prevalence of HTN (42.72%) than women (39.00%). Our ndings are comparable to those of the 2019 STEPS survey Nepal and 2016 Nepal DHS, where the prevalence of HTN was 40.91% and 32.6%, respectively [24]. In the present study and these two nationwide surveys, the prevalence of HTN increased with increasing age group. In this study, 57.6% of hypertensive patients did not know they had raised blood pressure levels before this study, whereas it was 72.5% of hypertensive adults of the same age group in the 2019 STEPS survey of Nepal.
The prevalence of obesity was higher in the present study than in the nationwide survey, but the prevalence of hypertension was similar. This study shows that screening for obesity and hypertension at the health care facility level can catch similar numbers of obese, hypertensive people than lengthy and costly community surveys. This can be a real game changer in the opportunistic screening of obesity and hypertension in lowand middle-income countries (LMICs). Opportunistic screening can lead to early diagnosis of and, if followed up by intervention for obesity and hypertension, will prevent many NCDs, which in turn can reduce DALYs, improve the quality of life of people and reduce the economic burden of those countries arising from needing to tackle NCDs.

Obesity and Hypertension
The growing prevalence of overweight and obesity has been increasingly recognized and well established as one of the most critical risk factors for the development of hypertension. In the present study, a signi cantly higher prevalence of hypertension was found among participants with either overweight or obesity compared to participants with a healthy weight for all metrics. Most studies have shown that an increase in BMI also raises blood pressure levels [55][56][57][58]. The 2016 Nepal DHS also showed that rates of hypertension among women and men increase with increasing BMI. In the present study, the participants with WC greater than the cutoff were nearly twice (1.81) likely to be hypertensive. A study in Italy in 2001 also showed similar results, where men were three times more likely, and women were twice likely to have hypertension with WC higher than the cutoff point [59]. A study in the southeastern US also showed that hypertension increased with increasing WC [60], whereas a study among adults aged 18 and above in Srinagar, India showed that 69.0% of people with abnormal WC were hypertensive [61]. In the present study, the participants with WHtR ≥ 0.5 had odds of 1.81 being hypertensive. A longitudinal study between 2005-2008 among Korean adults aged 39-72 years showed that the participants with the highest quartile of the WHtR (WHtR≥0.54) were 4.51 times more likely to have hypertension [62]. A cohort study in Brazil among adults aged 18 years and above also showed that participants with beyond optimal BMI, WC and WHtR were more hypertensive than within normal metrics, and 48.0%, 54.0%, and 73.0% of participants who were overweight at baseline according to BMI, WC and WHtR, respectively, developed HTN over time [63]. The present study corroborates the fact that overweight and obesity are signi cant risk factors for hypertension.
In the present study, we found a statistically signi cant positive correlation between all anthropometric metrics and both SBP and DBP.  [67].

Limitations of the study
Although prospective, this is still an observational, cross-sectional study design; hence, we cannot infer a causal relationship between an increase in weight over an optimal level and raised blood pressure. The higher value of this study lies in its ability to signal out that a very high proportion of people coming to the hospital have overweight/obesity and hypertension that is being missed on a day-to-day basis in a clinical setting.

Conclusions
The present study showed that waist circumference, which is easy to measure, was both correlated and had higher predictive capacity among WHtR and BMI and may play a major role in the future diagnosis of HTN in Nepali adults. Regardless of the anthropometric metrics used to measure overweight and obesity, the hospital setting is an opportunity centre to screen for overweight, obesity and hypertension, which are major risk factors for NCDs.

Declarations
Ethics approval and consent to participate Ethical approval for this study was obtained from the Nepal Health Research Council (Reference no. 2910).
Informed consent was obtained from every participant before the interview and anthropometric measurement.

Consent for publication
Not applicable Availability of data and materials The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

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

Funding
No funding received for this study Authors' contributions RS and MPU designed the study. RS, BK, and JRB were involved in proposal writing. RS and BK were involved in data analysis. RS, BK, and SKU were involved in drafting the manuscript. RS, BK, MPU, JRS, SKU and MK are involved in critical analysis and review of the manuscript. All authors have read the manuscript carefully and approved its submission.