Association of smoking with coronary artery disease in Nepalese populations: a case control study

There are well-known traditional risk factors for coronary artery disease (CAD).

Abstract

There are well-known traditional risk factors for coronary artery disease (CAD). Among them, smoking is one of the most prominent and modifiable risk factors. This study aims to determine the magnitude of smoking as a risk factor for CAD in the Nepalese population. A hospital-based age- and sex-matched case–control study was carried out with a total of 612 respondents. Bivariate analysis showed that the risk of developing CAD in ex-smokers and current smokers was higher (odds ratio (OR): 1.81 (confidence interval (CI): 1.21–2.7) and OR: 5.2 (CI: 3.4–7.97)), with p-values less than 0.004 and <0.00001, respectively, compared to the risk in never smokers. From stratified socio-demographic, cardio-metabolic, behavioural and psychosocial risk factor analysis, smoking was found to be associated with CAD in almost all subgroups. In the subsequent multivariate analysis, adjustment for socio-demographic, cardio-metabolic and psychosocial risk factors showed a steady increase in risk. However, further adjustment for behavioural risk factors (alcohol use and physical activity) showed that the risk was attenuated by 59% in current smokers. After adjusting for the covariates, current smokers and ex-smokers had an increased risk of CAD (OR: 6.64, 95% CI: 3.64–12.12, p < 0.00001; OR: 1.89, 95% CI: 1.08–3.31, p < 0.012, respectively) compared with non-smokers. In conclusion, smoking was found to increase the risk of CAD in the Nepalese population.

Introduction

Cardiovascular disease (CVD), cancer, chronic obstructive pulmonary disease (COPD) and diabetes have been identified by the World Health Organization (WHO) as the four major non-communicable diseases (NCDs) worldwide.1 Each year, 15 million people between the ages of 30 and 69 years die from a NCD, and over 80% of these “premature” deaths occur in low- and middle-income countries, demonstrating an important public health challenge.2 Tobacco use, unhealthy diet, lack of physical activity, and harmful alcohol use, are four key modifiable behavioural risk factors which contribute the most to the development of NCDs.1 Of these, age, sex, smoking, obesity, dyslipidaemia, physical inactivity, hypertension and diabetes mellitus are established risk factors for CVD.3,4 Most studies on the risk factors of coronary artery disease (CAD) have been reported from research carried out in Europe and North America, and sufficient information regarding the relative contribution of these risk factors in south Asian communities is still lacking.4

Tobacco use is the leading cause of death worldwide, killing 5 million people each year and accounting for 12% of all deaths.5 In addition, 29% of deaths from coronary heart disease can be attributed to tobacco.6 Smoking and passive smoking have been identified as variable risk factors for acute myocardial infarction (AMI) and CAD.7 In high-income countries, along with the decreasing prevalence of smoking, a decline in CAD, due to the decline in important risk factors, has been observed over the past few decades. Despite this improvement, smoking prevalence remains high in many of these countries, and no substantial effort has been made in low-income countries to reduce this prevalence.6 However, a number of studies have shown that smoking cessation significantly reduces CVD prevalence.8 The INTERHEART study4 was a case–control study conducted in 520 countries (Africa, Asia, Australia, the Middle East, and North and South America) which showed that smoking was one of the most important causes of global AMI, especially in men. Smokers have an increased risk of non-fatal myocardial infarction compared with non-smokers (odd ratio (OR): 2.95, 95% confidence interval (CI): 2.77–3.14, p < 0.0001), and their risk increased by 5.6% per cigarette. The risk associated with former smokers decreased to 1.87 times within 3 years after quitting smoking (95% CI: 1.55–2.24); however, there are still residual risks in the 20 years after smoking cessation (1.22, 1.09–1.37). In a large multi-ethnic population,9 the OR of AMI for smokers was 2.95, which was higher than that of non-smokers.

The burden of CVD is also increasingly affecting developing countries such as Nepal. Similar to other low- and middle-income countries, Nepal is facing a triple burden of diseases: communicable diseases, re-emerging diseases and an escalation of NCDs. Major NCDs, such as CVD, diabetes, cancer and COPD, are in increasing prevalence.10,11 As Nepal is a resource poor country, if there is a continuous increase in disease burden by non-communicable disease such as CAD, Nepalese people cannot afford the costly curative services.12 As in high-income countries, the major risk factors in the Nepalese population include tobacco use, alcohol consumption, unhealthy diet and physical inactivity,13,14 which were recognized based on a prevalence study. Moreover, although it has policy on smoking reduction such as banning smoking in public places and writing the risk of smoking on the cover of a cigarette box, the obvious risk of CAD has still not been documented.10 Therefore, the magnitude of association of smoking with CAD remains unknown in the Nepalese population. Knowing the risk of the disease, it is also better to allocate resources with better policy in a more effective and efficient way. The present study analyses smoking as a risk factor for the development of CAD in the Nepalese population.

Materials and methods

Study design and participants

A hospital-based case–control analytical study was designed to assess the strength of the association between smoking and CAD. In total, 306 cases and 306 control respondents were recruited between June and October 2017 from a highly specialized central level public cardiac treatment centre, Shahid Gangalal Cardiovascular National Heart Center, in Nepal. The hospital is a referral centre for people from all areas of the country regardless of their socioeconomic status. Thus, the study was carried out in this hospital because the general population from all areas of the country attends this facility for diagnosis and treatment.

An individual was considered a unit of analysis; therefore, hospitalized patients and patients who were admitted for angiography were included as cases. Hospitalized patients were those patients who were admitted after having suspected myocardial infarction (MI), cardiac bypass surgery (CABG), or angioplasty. The angiographic criteria were MI or stress test-positive patients or diabetic patients experiencing chest pain. If stenosis was greater than 70% based on angiography,15 the patient was eligible for inclusion in this study as a case. Patients were interviewed on the second or third day after angiography if they were stable. Only acute and non-fatal cases of CAD were included in the study. Regarding angioplasty and CABG patients, those who were diagnosed with acute CAD and had records for the following 2 months after angiography were included. Moreover, old MI cases, patients in the intensive coronary unit and severely ill patients, such as those with kidney failure and heart failure, were excluded from this study. Controls were selected from those patients who came for general check-up at the outpatient department. To determine if patients were eligible for the treadmill exercise test (TMT), cardiologists screened them on the basis of their chief complaint, such as chest pain and/or with a history of hypertension, diabetes and dyslipidaemia. Additionally, a whole-body check-up including the TMT was performed upon requests from patients. If the patient had a negative TMT result, he/she was eligible for inclusion in the control group. Additionally, if patients presented normal angiography results, they were eligible for inclusion in the control group in this study.

Data collection

Face-to-face interviews and observations were performed using a semi-structured questionnaire and observation checklist to obtain reliable information from respondents (both cases and controls). Blood pressure, weight and height, and waist and hip circumference measurements of all cases and controls were obtained, and the results were recorded using an observation checklist. Validity and reliability testing of the questionnaire were performed before conducting the study.

A standardized protocol was used to measure the height, weight, and waist and hip circumferences. The height was measured using a wall-mounted stadiometer to the nearest centimetre. We asked respondents to stand upright without shoes, with their back against the wall, heels together and eyes directed forward. Their weight was measured with a portable electronic weighing scale placed on a firm horizontal surface. Waist and hip circumferences were measured with a non-stretchable standard measuring tape. Waist measurements were obtained over a lightly clothed abdomen at the narrowest point between the costal margin and iliac crest, and the hip circumference was measured over light clothing at the level of the widest diameter around the buttocks. The body mass index (BMI) was categorized as normal (<25.0 kg m^–2), overweight (25.0– < 30.0 kg m^–2), and obese (≥30.0 kg m^–2). Abdominal obesity was defined as a waist circumference ≥90 centimetres in males and ≥80 centimetres in females.16 Dyslipidaemia was defined as hypercholesterolemia or hypertriglyceridemia; if the high-density lipoprotein (HDL) level was below 30 g dl^–1, dyslipidaemia was considered. Hypertriglyceridemia and hypercholesterolemia were defined as triglyceride (TG) serum and total cholesterol (TC) levels greater than 150 and 200 mg dl^–1, respectively, or if hypolipidaemic treatment was administered. Diabetic individuals were those with a fasting blood glucose level equal to or greater than 126 mg dl^–1 over the course of two tests or those who were taking diabetes medications.17 Patients whose blood pressure values were greater than or equal to 140/90 mmHg or those who were taking antihypertensive medication according to their medical records were classified as hypertensive.18

Persons who smoked or used smokeless tobacco daily until within 1 year of the interview were considered current daily smokers and daily tobacco users. Never smokers were those who responded with “occasionally” or “not at all” on the questionnaire, and ex-smokers were those who smoked daily prior to 1 year of interview. Ex-smokers and current smokers were asked about the age at which they began smoking; similarly, ex-smokers were asked about the time when they quit smoking. Additionally, current and ex-smokers were asked the number of cigarettes usually smoked per day. To determine the association between smoking and CAD, the never smoker group was considered the reference group. Current alcohol drinkers were categorized as those who engaged in alcohol drinking within the last year. A total of 13.6 g of pure alcohol was considered as one standard drink, which was equivalent to consuming 43 ml of local alcohol (Raksi) and 341 ml of beer, Zaand or Tongba.19

Physical activity levels related to work were categorized as vigorous, moderate or low. Vigorous physical activity was considered any activity that caused a substantial rise in heart and breathing rates, for example, digging or ploughing fields, lifting heavy weights, etc. Continuous engagement in such activity for at least 10 minutes was considered involvement in vigorous activity. Similarly, moderate physical activity was defined as any activity that caused a moderate increase in heart and breathing rates (examples include domestic chores, gardening, lifting light weights, etc.). Continuously engaging in such activity for at least 30 minutes was considered involvement in moderate activity. Physical activity related to transport was not considered in this study. Recreational activity, also called physical exercise, included two types of activities, vigorous and moderate, based on exertion. Vigorous recreational activity was defined as any recreational activity that causes a large increase in heart and breathing rates, for example, activities such as football, fast swimming and rapid cycling. Ten minutes of such activity was considered involvement in vigorous recreational activity. Moderate recreational activity was defined as any kind of recreational activity that causes a moderate increase in heart and breathing rates; examples include yoga, playing basketball, brisk walking, and normal cycling. Sedentary behaviour was defined as a behaviour in which an individual spends time sitting at a desk; sitting with friends; travelling in a car, bus or train; reading a book; and so on.20

Data analyses

Data were entered in Excel and later analysed using SPSS version 22.0 (SPSS Inc., Chicago, IL, USA). The distribution of the data was tested for normality to determine the use of parametric or non-parametric tests. Categorical variables are presented as the frequency and percentage, whereas continuous variables are presented as the mean ± standard deviation. For bivariate analyses, differences in the means were detected using two sample independent samples t-test or Wilcoxon tests, as appropriate, and differences in proportions were detected with the chi-square test. We used an ordinary logistic regression analysis to determine the strength of the association between smoking and the development of CAD, with the development of CAD as the dependent variable, and ORs and 95% CIs were calculated. Two-tailed p-values <0.05 were considered significant.

Multivariable adjusted models were constructed with incremental adjustments. The first model was an age group and gender adjusted model. In model 2, we further adjusted for factors (age group, sex, ethnicity, occupation, education and family income). In model 3, further adjustment for cardio-metabolic risk factors was performed to observe the effects of hypertension, diabetes, dyslipidaemia, and central obesity (weight to height (WH) ratio). In model 4, further adjustment for psychosocial factors (mental illness, sleep disturbance and sleep hours) and family history was performed, in addition to the factors in model 3. Adjustment for socio-demographic factors, cardio-metabolic risk factors, psychosocial factors, family history and behavioural risk factors (physical activity and alcohol use) was performed in the last model (fifth model). All the models were also assessed using the Hosmer and Lemeshow goodness-of-fit tests.

Ethical approval

Before conducting the survey, permission for clinical data collection was obtained from the respective hospital. Ethical approval was obtained from the Ethical Review Board of the Nepal Health Research Council (Reg. no. 308/2018). Verbal informed consent was obtained from the respondents before conducting the interviews.

Results

Table 1 shows the socio-demographic characteristics of all respondents. The mean age of the patients was 56.35 ± 10.5 years. Stratification by sex showed that there was a significant association between smoking and CAD in both males and females. Regarding ethnicity, smoking was associated with CAD in the Brahmin/Chetri, Janajati, and Terai castes but not in the Newar and Dalit castes. Smoking was also highly significantly associated with CAD in the agriculture and labour groups and in the education and family income categories.

Table 1. Socio-demographic characteristics of respondents.

Variables	Control (%) N = 306				Case (%) N = 306
	Total	Non smoker	Ex-smoker	Current smoker	Non smoker	Ex-smoker	Current smoker	P-value
Age	612	55.32 ± 10.36	61.03 ± 9.88	54.43 ± 11.17	56 ± 11.12	59.32 ± 9.26	54.86 ± 10.34
Gender
Female	146(23.9)	58(79.5)	13(17.8)	2(2.7)	45(61.6)	12(16.4)	16(21.9)	0.002
Male	466(76.1)	138(59.2)	57(24.5)	38(16.3)	68(29.2)	61(26.2)	104(44.6)	0.0001
Marital status
Married and living with spouse	576(94.1)	190(64.4)	66(22.4)	39(13.2)	100(35.6)	70(24.9)	111(39.5)	0.0001
Unmarried/divorced/separated/lost spouse	36(5.9)	6(54.5)	4(36.4	1(9.1)	13(52)	3(12)	9(36)	0.115
Ethnicity
Brahmin/Chhetri	301(49.2)	100(61.3)	41(25.2)	22(13.5)	38(27.5)	44(31.9)	56(40.6)	0.0001
Newar	96(15.7)	25(55.6)	13(28.9)	7(15.6)	27(52.9)	8(15.7)	16(31.4)	0.109
Janajati	70(11.4)	24(68.6)	7(20)	4(11.4)	11(31.4	8(22.9)	16(45.7)	0.002
Dalit	33(5.4)	10(58.8)	2(11.8)	5(29.4)	37(80.4)	7(15.2)	2(4.3)	0.05
Chaudhari/Terai castes	112(18.3)	3(18.8)	2(12.5)	11(68.8)	34(51.5)	11(16.7)	21(31.8)	0.001
Occupation
Household work	91(14.9)	37(72.5)	11(21.6)	2(5.9)	23(57.5)	6(15)	11(27.5)	0.017
Self-employed/entrepreneurs	113(18.5)	54(79.4	6(8.8)	8(11.8)	15(33.3)	9(20)	21(46.7)	0.001
Regular employee of an organization	137(22.4)	38(60.3)	14(22.2)	11(17.5)	31(41.9)	15(20.3)	28(37.8)	0.026
Agriculture/labor	139(22.7)	33(56.9)	14(24.1)	11(19)	19(23.5)	19(23.5)	43(53.1)	0.0001
Retired/unemployed	132(21.6)	34(51.5)	25(37.9)	7(10.6)	25(37.9)	24(36.4)	17(25.8)	0.062
Education
Illiterate	141(23)	32(51.6)	21(33.9)	9(14.5)	30(38)	14(17.7)	35(44.3)	0.001
Primary school certificate	167(27.3)	45(59.2)	19(25)	12(15.8)	23(25.3	24(26.4)	44(48.4)	0.0001
Middle/high school certificate	143(23.4)	49(72.1)	10(14.7)	9913.2)	35(46.7)	15(20)	25(33.3)	0.005
Intermediate/diploma	57(9.3)	21(63.6)	8(24.2)	4(12.1)	11(45.8)	9(37.5)	4(16.7)	0.405
Graduate/profession/honors	104(17)	49(73.1)	12(17.9)	6(9)	14(37.8)	11(29.7)	12(32.4)	0.001

Table 2 shows the association of smoking across cardio-metabolic risk factor groups. Diabetes, hypothyroidism and obesity were strongly associated with CAD, whereas hypertension was not associated with CAD eventhough it is an established risk factor for CAD. Crucially, analysis of the presence or absence of these cardio-metabolic risk factors showed that smoking was associated with CAD in all categories, except in the diabetic and underweight populations.

Table 2. Association of smoking across cardio-metabolic risk factor groups.

Variables	Control (%) N = 306				Case (%) N = 306
	Total	Non smoker	Ex-smoker	Current smoker	Non-smoker	Ex-smoker	Current smoker	P-value
Diabetes T2DM
No	486(79.4)	171(64.3)	59(22.2)	36(13.5)	73(33.2)	43(19.5)	104(47.3)	0.0001
Yes	126(20.6)	25(62.5)	11(27.5)	4(10)	40(46.5)	30(34.9)	16(18.6)	0.214
Hypertension
No	327(53.4)	105(64.4	31(19)	27(16.6)	53(32.3)	37(22.6)	74(45.1)	0.0001
Yes	285(46.6)	91(63.6)	39(27.3)	13(9.1)	60(42.3)	36(25.4)	46(32.4)	0.0001
Dyslipidemia
No	490(80.1)	175(63.9)	66(24.1)	33(12)	77(35.6)	51(23.6)	88(40.7)	0.0001
Yes	122(19.9)	21(65.6)	4(12.5)	7(21.9)	36(40)	22(24.4)	32(35.6)	0.043
Obesity
Underweight	24(3.9)	2(22.2)	4(44.4)	3(33.3)	2(13.3)	4(26.7)	9(60)	0.449
Normal	273(44.6)	99(62.7)	35(22.2)	24(15.2)	42(36.5)	29(25.2)	44(38.30	0.0001
Overweight	250(40.8)	77(68.1)	23(20.4)	13(11.5)	51(37.2	30(21.9)	56(40.9)	0.0001
Obese	65(10.6)	18(69.2)	8(30.8)	0	18(46.2)	10(25.6)	11(28.2)	0.011
Waist–hip ratio (WHR)
Normal	206(33.7)	89(62.7)	30(21.1)	23(16.2)	11(17.2)	17(26.6)	36(56.3)	0.0001
Excess	406(66.3)	107(65.2)	40(24.4)	17(10.4)	102(42.1)	56(23.1)	84(34.7)	0.0001
Waist circumference
Normal	193(31.5)	74(61.2)	27(22.3)	20(16.5)	19(26.4)	18(25)	35(48.6)	0.0001
Excess	419(68.5)	122(65.9)	43(23.2)	20(10.8)	94(40.2)	55(23.5)	85(36.3)	0.0001

Table 3 shows behavioural risk factors of CAD. Smoking, tobacco use, alcohol consumption and physical activity were associated with CAD. A total of 120 (39.2%) CAD patients were current smokers, whereas only 40 (13.1%) patients in the control group were current smokers. However, only a few more ex-smokers (73 (23.9%)) were observed in the CAD group than those in the control (70 (22.9%)) group. The median duration for quitting was 12 months (Q1, Q3: 6, 20) and 7.5 months (Q1, Q3: 7, 31.75) for cases and controls, respectively. Similarly, the median duration of total smoking was 34 months (Q1, Q3: 23, 44) and 4.5 months (Q1, Q3: 1, 28) for cases and controls, respectively. Both the median duration of quitting and total duration of smoking were significantly associated with CAD; however, the number of cigarettes was not associated with CAD. Regarding physical activity, leisure physical activity was associated with CAD, whereas work-related physical activity was not associated with CAD.

Table 3. Distribution of behavioral risk factors among cases and controls.

Variables	Categories	Cases (%) N = 306	Controls (%) N = 306	P-value
Drinking alcohol	Yes	91(29.7)	66(21.6)	0.013
Drinking alcohol as per amount
	No	212(69.3)	241(78.8)	0.007
	Moderate	18(5.9)	20(6.5)
	Heavy	76(24.8)	45(14.7)
Frequency of alcohol use/week		3(1, 4)	4(3, 5)	0.001
Amount of alcohol use per drink		6(5.75, 8.25)	6(4, 8)	0.075
Smoking
	Never	113(36.9)	196(64.1)	0.000
	Ex-smoker	73(23.9)	70(22.9)
	Current smoker	120(39.2)	40(13.1)
Average number of sticks per day (for a current smoker)		12(6, 21)	8(4.25, 19.5)	0.185
Time of initiation		17(15, 21)	18(15, 23)	0.137
Duration of smoking		34(23, 44)	4.5(1, 28)	0.000
Duration of quitting (for an ex-smoker)		12(6, 20)	17.5(7, 31.75)	0.020
Tobacco consumption	Current user	54(17.6)	37(12.1)	0.034
Time of initiation of tobacco consumption		21(18, 24)	21(18, 25)	0.886
Duration of tobacco consumption		30(17, 38)	22(12, 33)	0.011
Physical activity related to work (PARW)	Moderate/hard	172(56.2)	165(53.9)	0.569
Leisure time physical activity (LTPA)	Moderate/hard	68(22.2)	113(36.9)	0.000
Total physical activity (PARW + LTPA)	Moderate/hard	202(66)	241(78.8)	0.000
Duration of total rest per day
	<5 h	49(16)	36(11.8)	0.000
	5–10 h	133(43.5)	208(68.2)
	>10 h	124(40.5)	61(20)

Fig. 1 presents smoking as a risk factor compared to other potential risk factors in the multivariate analysis; almost all variables were found to be associated with CAD in the bivariate analysis (<0.05). The major risk factor for CAD was sleeping disorder. Compared to non-smokers, the risk of CAD increased 7.02-fold (95% CI, 4.17–11.87), with a p-value <0.00001, in current smokers and 2.53-fold (95% CI, 1.43–4.48), with a p-value <0.012, in ex-smokers.

Fig. 1. Smoking as a risk factor for CAD compared to other major risk factors (socio-demographic, cardio-metabolic, psychosocial and behavioural risk factors).

Table 4 shows the effects of smoking on CAD in the 5 different models. Model 1 was an age group and gender adjusted model in which the risks of ex-smokers and current smokers developing CAD exhibited OR values almost 1.81 times (95% CI: 1.21–2.7, p < 0.004) and 5.2 times (95% CI: 3.4–7.97, p < 0.00001) higher than those of never smokers. In the second model, adjustment for socio-demographic factors, namely, age group, sex, ethnicity, occupation, education and family income, was performed; ORs, 95% CIs and p-values for ex-smokers and current smokers were 2.08; 1.321, 3.284, and <0.002 and 6.26; 3.843, 10.183, and <0.00001, respectively. Further adjustment for cardio-metabolic risk factors (hypertension, diabetes, dyslipidaemia, and central obesity) showed an additional increase in risk in current smokers (6.92; 4.011, 11.935), with a p-value <0.00001, but a slight decrease in risk in ex-smokers (1.89; 1.151, 3.118), with a p-value <0.012. In model 4, further adjustment for mental health related risk factors (mental illness, sleep disturbance and sleep hours) and family history was performed. Similar results to those in model 3 were observed in both groups; in the current smoker group (7.23; 4.016, 13.031), the p-value was <0.00001, but there was a slight decrease in risk in the ex-smoker group (1.89; 1.151, 3.118), with a p-value <0.00001. In the last model (model 5), further adjustment for behavioural risk factors (alcohol use and physical activity) was conducted. The risk was attenuated by 59% in smokers, but was still similar to model 4.

Table 4. Association between smoking and coronary artery disease in the Nepalese population.

Model		Never	Ex-smoker	Current
Model 1	OR	Ref.	1.81	5.20
	95% CI		1.21–2.7	3.4–7.97
	P-value		<0.004	<0.00001
Model 2	AOR	Ref.	2.08	6.26
	95% CI		1.321–3.284	3.843–10.183
	P-value		<0.002	<0.00001
Model 3	AOR	Ref.	1.89	6.92
	95% CI		1.151–3.118	4.011–11.935
	P-value		<0.012	0.00001
Model 4	AOR	Ref.	1.87	7.23
	95% CI		1.08–3.25	4.016–13.031
	P-value		<0.00001	<0.00001
Model 5	AOR	Ref.	1.89	6.64
	95% CI		1.082–3.31	3.642–12.105
	P-value		0.012	0.00

Fig. 2 shows that in current smokers, there was a consistent increase in the risk ratio from models 1 to 4 but not in model 5.

Fig. 2. Association between current smokers and CAD across the models (models subsequently adjusted for age group and gender, and socio-demographic, cardio-metabolic, psychosocial and behavioural risk factors).

Fig. 3 shows that in ex-smokers, the risk ratio increased in model 2 but remained constant across the remaining models. Unlike in the current smoker group, no further decrease in risk was observed in model 5 when physical activity was added in model 4.

Fig. 3. Association between ex-smokers and CAD across the models (models subsequently adjusted for age group and gender, and socio-demographic, cardio-metabolic, psychosocial and behavioural risk factors). CAD = coronary artery disease; OR = odds ratio; AOR = adjusted odds ratio; Ref. = reference category.

Discussion

In the present study, we considered socio-demographic, cardio-metabolic, behavioural and psychosocial risk factors. As the study was age group and sex matched, in bivariate analysis, among the risk factors, age and sex were not associated with CAD because of their similar values in both case and control groups. Although hypertension is a well-established risk factor among the cardio-metabolic risk factors,21 there was no significant association with CAD because of the selection of control patients with a high prevalence of hypertension. In some studies, low and moderate consumption of alcohol was found to be a protective factor for CAD,17 but we found discrepant results because we did not categorize alcohol use as heavy or moderate. In addition to these two variables, as in most other studies,4,22 occupation, diabetes mellitus, dyslipidaemia, obesity, family income, family history, physical activity, mental illness, and sleep disturbance were found to be significantly associated with CAD. In this study, there was a high prevalence of current smokers (39.2%) and ex-smokers (23.9%) presenting CAD. A similar prevalence was also documented by Adhikari et al. in a Nepalese population.23 However they conducted the cross-sectional study with a sample size of 495 ST elevation myocardial infarction patients. In a Pakistani population, smoking (29.2%), predominantly in males, has been reported among patients with CAD.24

Although the risk of developing CAD was higher in males than in females,4 female smokers were more prone to CAD than their male counterparts.25 We found that current smoking and ex-smoking were related to CAD in both sexes. Additionally, in a Chinese population, current female smokers had a greater risk of heart attack than non-smokers and ex-smokers.26 In this study, the mean age of patients was 56.35 ± 10.5 years. When the age was categorized, patients aged <50 years had a higher chance of developing CAD. South Asian populations, especially Indian populations, in the young age group were more susceptible to CAD, with a prevalence of 5% to 10%. The most commonly associated risk factor in young CAD patients was smoking.27 In a hospital-based observational study, more smokers were in the CAD group (OR, 1.59; 95% CI, 1.22–2.07; p = 0.001) with young adults suspected of coronary artery disease at ≤35 years of age.28 Traditional risk factors are strong predictors of the incidence of CAD, but their association with disease could differ across ethnic groups.29 In our study, in all ethnic groups, except in Newar and Dalit, smoking was found to be associated with CAD; further genetic studies might be helpful in determining the reason behind this phenomenon. In the stratification by occupation, a significant association between smoking and CAD was observed in each occupational group. Smoking was inversely related to education among males but not females (p for interaction between sex and smoking = 0.005).30 Additionally, in the present study, the dfferent results of the stratification by education in all groups may be due to an inadequate sample size. A family history of coronary heart disease adjusted for age, sex, smoking, and the geographic region produced an OR of 1.55 (99% CI: 1.44–1.67).4 In our study, smoking was also found to be associated with a family history of CAD.

When stratified by alcohol use as “yes” or “no”, smoking seemed to be strongly associated with CAD in the present study. Moderate alcohol consumption decreased the risk of CAD severity compared with heavy drinking in a study conducted in patients undergoing angiography.17 In this study, smoking was related to CAD in both physically active and inactive patients. During a follow up of two years in CAD patients with T2D, an inverse association between habitual LTPA and a short-term cardiovascular outcome was observed, but controlled, home-based exercise training showed only minor effects on the CV risk profile.31 However, when 25-year physical activity (PA) trajectories were evaluated, those people who did physical activity three times higher than the recommended level had a high chance of developing coronary subclinical atherosclerosis by their middle age.32

Smokers have an increased risk of coronary heart disease, which is dependent on the number of cigarettes smoked daily and the number of years they have smoked.33 In this study, the number of average cigarettes per day, total duration of smoking and length of quitting were associated with CAD risk. Consistent with this study, current smokers who smoked more than 10 cigarettes per day were found to have a significantly increased risk of CAD. The findings of that study also strongly suggested that early smoking cessation could possibly prevent future cardiac events.34

Initially, adjustment was performed for socio-demographic factors, namely, the age group, sex, ethnicity, occupation and education. There was an increase in the CAD risk in both groups after adjusting for socio-demographic risk factors. Further adjustment for cardio-metabolic risk factors (hypertension, diabetes, dyslipidaemia, and central obesity) showed an additional increased risk in current smokers but a slightly decreased risk in ex-smokers. Smoking, hypertension, and the metabolic syndrome (MetS) were significantly more common among patients with coronary heart disease.35 In model 4, further adjustment for mental health risk factors (mental illness, sleep disturbance and sleep hours) and family history was performed. In this model, the chance of current smokers developing CAD was similar to that of the previous model, but the chance in ex-smokers was slightly decreased. In the last model, data were further adjusted for behavioural risk factors (alcohol use and physical activity). The risk was attenuated by 59% in smokers, but it was still similar to that in model 4 because of physical activity, which was notably observed as a protective factor for CAD in the present study. The risk of developing CAD for ex-smokers and current smokers was still almost 2.5-fold and 7-fold higher, respectively, than that of non-smokers. CAD risk results vary country to country because the prevalence of smoking varies. A similar result has been reported in a previous study, where the age-standardized OR of ST-elevation myocardial infarction (STEMI) was 5.2 (CI: 4.5–6.1) for current smokers and 1.1 (CI: 1.0–1.3) for ex-smokers compared to those of never smokers,8 whereas only double the risk was observed in the Spanish population.36 In one study of non-smokers and current or previous smokers in Jordan, 56.7% of CAD patients were smokers and smoking doubled the risk of CAD compared to not smoking.22 A significant association of past-year smoking was with CAD reported from patients hospitalized with coronary artery disease in an American population.37 In a recent study among the Han population in China, gene–environment interactions between rs4886605 and smoking increased the risk of CAD.38

The present study is the first study conducted in the Nepalese population to quantify the association between smoking and CAD. There were some limitations to this study. The sample size was not too large; therefore all variables could not be adjusted in the models, and the study included only non-fatal cases. Questionnaire evaluation of smoking cannot eliminate the impact of second-hand smoke. Serological objective indicators (e.g. serum cotinine) are more accurate. Naturally, the case control design was influenced by a lot of bias which caused difficulty in arriving at any firm conclusions about causality. In our study selection, measurement and recall bias pose difficulty in making the result valid. Therefore further cohort study is needed to make the result more generalizable. However, we tried to moderate the effect of bias to make the results valid. Firstly a matched case–control study was carried out to minimize confounding and maximize comparability. Secondly our results were well in line with studies from elsewhere, and we believe that the results are valid. Face to face interviews were conducted to take the information accurately and to minimize the recall bias. Although physical activity and smoking behaviour were self-reported and therefore were more likely to be affected by a random measurement error, the height and weight were measured objectively. For elimination of measurement bias, repeated measurements with standard tools were performed. Selection bias was minimized by including controls from the same hospital.

Conclusion

The present study confirmed that smoking was an important risk factor for CAD in the Nepalese population. Even after stratification by socio-demographic, cardio-metabolic, behavioural and psychosocial risk factor groups, smoking seemed to be associated with CAD. Study replication and further applied research are required to estimate the interaction between smoking and other traditional and modern risk factors. A serological study can well predict the effect of smoking on CAD.

Conflicts of interest

There are no conflicts to declare.