This was a descriptive, community based, cross-sectional study conducted between January and April 2014 in Tinkanya Village Development Committee, Sindhuli, Nepal. Study site is an extremely hilly, remote, rural community where mostly indigenous people live. Developmental markers like roads, electricity and health facilities still remain almost nonexistence. A part from livestock and farming, remittance is the major source of livelihood. Most working age men have migrated to city/abroad for better employment.

The community contains 805 households including 4169 population residing in nine Wards [19]. We divided nine Wards into three clusters and organized a screening camp on each clusters separately. Community leaders, teachers and elderly persons were mobilized for encouraging eligible people for their voluntarily participation.

We calculated the sample size using one sample situation of estimating population proportion with specified absolute precision formula [20]. The estimation of population prevalence of risk factors was considered as 50% with equal percent (5%) of allowable error, level of significance and non response rate. Based on predefined sample size and sampling method, we included 406 participants, almost equal numbers from each cluster, who provided informed consent to involve in study. Every third registered participants who met the inclusion criteria of having age between 20 and 50 years and without an established CVD were enrolled for the study.

Study protocol was reviewed and approved by Institutional Review Board of Institute of Medicine, Kathmandu. We obtained the verbal consent from all participants after detailed explanation of research purpose, and assurance of maintaining privacy and confidentiality. Free heath checkup and medicines were provided to the participants based on their health problems. Those who needed further treatment were referred to higher centers. A successive awareness camp was also organized for educating people on CVD risk factors.

We followed the WHO-NCD STEPwise approach to surveillance questionnaires for collecting demographic information, behavioural and anthropometric measurements [21]. It contains core and extended questionnaires related to smoking, alcohol consumption, fruits and vegetables intake, and physical activities [22]. STEP 2 core provides the measure of height, weight, waist circumference and blood pressure. We also added some structured questionnaires relevant to CVD risk factors into this tool.

All data enumerators were trained in research protocol, interviewing and measuring the variables. They were supervised by field co-coordinators experienced in national NCD risk factors surveillance. Research tool was appropriately modified after undertaking a pretest.

Weight was measured using CAMRY weighing machine. Height and waist circumference were noted in centimeter using Johnson’s non-stretchable measuring tapes. Waist circumference was measured in between the lower rib and anterior superior iliac crest. Doctor’s Aneroid Sphygmomanometer (BP Set) was used for recording blood pressure. We recorded three readings of systolic and diastolic blood pressure in five minutes interval over right arm and the latter two were averaged for final score. Socio-economic status was assessed using revised Kuppuswamy’s scale [23]. Average salt intake per person per day was calculated in four steps. First, the total days that one family could sufficiently use one kilogram of salt (one packet) were counted. If same salt was used for cattle, that amount was also roughly estimated and reduced from it. Then, the total salt consumption was divided by total days it was used and number of family members who were sharing the same kitchen in order to derive salt intake per person per day. Finally, it was averaged to calculate average salt intake per capita per day.

Standard operational definitions were adopted for behavioral, anthropometric and clinical measurements to ensure uniformity and minimize error. Smoking and alcohol consumption were considered as present if participants had had habit of taking them until the past 30 days of interview. Current smokers were defined as those reported smoking any tobacco product within last 30 days. Person smoking daily was regarded as current daily smoker. Similarly, current alcohol drinkers were labeled for those who engaged in alcohol drinking within last 30 days. In the same way, current episodic heavy drinking was considered as five and more drinks on any day in past 30 days. One standard drink (13.6 gm of pure alcohol) was equivalent to consuming 341 ml of beer or Zaand or Tongba; and 43 ml of local Raksi [9].

Fruits and vegetables intake for at least five portions (400 gm) of a day was regarded as sufficient [24]. One serving of vegetable was considered to be one cup (250 ml) of raw green leafy vegetables and ½ cup (125 ml) of cooked or chopped raw vegetables. One whole medium sized fruits like apple, banana or orange and ½ cup of chopped, cooked, canned fruit or ½ cup of fruit juice was in general considered as one portion [24]. Amount was estimated by using show cards. Sufficient physical activity was defined as at least 150 minutes moderate physical exercise in a week. Body mass index (BMI) was calculated and categorized as normal (<25.0 kg/m²), overweight (25.0– < 30.0 kg/m²), and obese (≥30.0 kg/m²) [25, 26]. Based on Asian criteria, abdominal obesity was defined as waist circumference ≥ 90 centimeters in men and ≥ 80 centimeters in women [27]. The diagnostic criterion for hypertension was set systolic blood pressure ≥ 140 mmHg and/or a diastolic blood pressure ≥ 90 mmHg as recommended by Joint National Committee-VII [28]. The persons who were using antihypertensive medicine were also listed under the category of hypertension.

Data were compiled, edited and checked to maintain consistency. Repetitions and omissions of data were corrected before coding and entering them in Epidata V.2.1. Recorded data were, then, exported to SPSS V.16.0 for further analysis. R i386 2.15.3 was used for comparing proportions and calculating confidence interval.

Descriptive statistics was used to identify the distribution of risk factors among socio-demographic characteristics. Chi-square and independent t test were conducted for comparing proportions of categorical and mean of continuous variables. All tests were two-tailed and p < 0.05 was considered statistically significant. Bivarate and multivariate analyses were conducted by binary logistic regression. Categorical variables were coded appropriately before entering them into model. Hypertension was considered as a dependent variable with dichotomous outcomes: 0 = Absent, 1 = Present. Other variables were coded as:- a) smoking: 0 = No, 1 = Yes, b) alcohol consumption: 0 = No, 1 = Yes, c) fruits and vegetables intake: 0 = sufficient, 1 = insufficient, d) physical activity: 0 = sufficient, 1 = insufficient.

Separate bivariate logistic regression analyses were conducted for independent variables. Correlation matrix was calculated to investigate their interrelationship. Variables those were identified as significant predictors in bivariate analyses were entered in multivariate model. Independent variables having significant correlation with each other were further investigated for interaction in multivariate analysis. The final multivariable model included age, smoking, alcohol consumption, salt intake and BMI. Waist circumference was dropped out from the multivariable model because of collinearity with BMI.

This study contained total 406 participants, including 176 (43.3%) male and 230 (56.6%) female. The mean age was 36.2 ± 9 years, with statistically non-significant (t = 2.086, df = 404, p = 0.38) mean difference of 1.879 years between male and female. Majority of participants belonged to Adhiwasi/Janajati (indigenous group) (63.1%), Hindu (80.3%), and were married (86.7) (Table 1). Most respondents (46.3%) had not been to school, were involved in household activities (62.6%) and had lower socio-economic status (76.3%) (Table 1).

Among total participants, 28.6% people were smoking until last 30 days of interview and 27.1% of people were smoking daily. Smoking rate was significantly higher among 41–50 age-group, married, illiterate (no formal education) and lower socio-economic status than others (Table 2). Among smokers 64.6% people started smoking before 15 years of age and 63.8% did not even think to give it up. Nevertheless, almost all (91.6%, 406) participants unanimously replied that smoking is injurious to health. Nearly 2/3 respondents (70 .4%) believed smoking and second hand smoke caused respiratory and cardiac problems.

Total 194 (47.8%) participants had habit of drinking alcohol. Among them, 14.3% people were taking alcohol daily. Comparatively people without formal schooling, of age 41–50 years, from indigenous group and having low socio-economic status had significantly high indulgence to alcohol than others (Table 2). The average pure alcohol consumption was 110 gm with maximum limit to 400 gm. Among alcohol users, 62.2% (25.1% in total respondents) consumed equal or more than five standard drinks at any day during last month.

Sufficient fruits and vegetables consumption (≥5 servings/day) was lacking in 96.6% of respondents. Mean number of days of fruits and vegetables consumed per week were 2.4 and 4.3 days respectively. In the same way, mean number of fruits and vegetables servings per day were 0.48 and 1.05. Mostly illiterate (no formal education), socially adjudged higher caste group (Brahman and Chhetri) and upper lower class people had significantly high prevalence of inadequate amount of fruits and vegetables consumption (Table 2). More than two third (75.9%) of respondents were using vegetables oil (mustard, soybean, sunflower oil) for cooking purpose. But, small proportion of people (2%) occasionally utilized animal fat in replacement of vegetable cooking oil. Daily salt intake per person in an average was 14.4 gm ±4.89 gm. Large number of people (11.2%) were still beyond the coverage of iodized salt. They were using un-iodized raw salt. Though the study population was residing in hilly terrain, nearly half of respondents (48.8%) insufficiently (≤5 days per week) involved in moderate physical activity. People with age between 20 and 30 years, and having higher education were involving in significantly less physical activity. Though there was no significant difference in physical activity level by gender and socio-economic status, comparatively higher percentage of male and upper middle class people had low physical activity than others (Table 2). Mean body mass index and waist circumference were 24.5 kg/m² and 79.8 cm respectively. Overweight and obesity (BMI ≥ 25 kg/m²) was present in 37.4% of people. It was relatively high in 31–40 age group and lower class population (Table 2). Over weight and obesity, separately, were observed in 29% and 8.4% of respondents. Abdominal obesity marker-waist circumference was above normal cut off point (≥80 cm in female, ≥90 cm in male) in 25.6% of participants.

The estimated prevalence of hypertension was 12.3% (9.1-15.5%). It had inverse relationship with education and socio-economic status. Hypertension was relatively higher in low socio-economic level, people without any formal education, married, male and the oldest age group than others. In bivarate analysis, hypertension was significantly associated with age (β = .08, df = 1, p = .000), smoking (β = 1.164, df = 1, p = .000), alcohol consumption (β = .75, df = 1, p = .016), salt intake((β = .097, df = 1, p = .004), and BMI (β = .112, df = 1, p = .006). These variables were entered in binary logistic regression model for multivariate analysis. All four variables, except alcohol consumption were identified as significant predictors of hypertension (Table 3). Hosmer and Lemeshow Test for goodness of fit was not statistically significant (p = 0.106) with fairly large pseudo R square (Nagelkerke R Square = 0.222).

Multivariate analysis underscored smoking as a strong predictor of hypertension. Smoker was 2.285 times more likely to have hypertension than non-smoker (Table 3). Following smoking, salt intake remained to be second strong explanatory variables for hypertension. Every one gram more intake of salt increased the probability of being hypertensive by almost 14%. Similarly, age was also positively associated with hypertension (Table 3). If person was getting older, there was more likelihood of getting hypertension. Excessive fat deposition on body also had high tendency to increase blood pressure. One unit addition in body mass index could heighten the risk by 0.088 times for being hypertensive, holding other predictors constant.

Significantly high numbers of female and indigenous group participants were involved in study. However, it was unintentional and was true representation of the present community where mainly women and indigenous people were living. Study findings should cautiously be generalized to other remote rural communities. Method used for measuring daily salt intake per person could be inaccurate. A more accurate way to determine dietary sodium, such as 24-hour urine collection, is warranted to confirm our findings. Similarly, scale used for determining socio-economic status has not yet been validated in Nepal.