This article is mentioned in:

Introduction

Smoking is one of the major lifestyle factors that influences the health of human beings (1). It is not only a risk factor for a multitude of diseases, but also one of the leading causes of avoidable mortality worldwide. Cigarette smoking is increasing rapidly throughout the developing world (2). Among the industrialized countries, smoking is estimated to be the primary cause of 70–90% of lung cancer cases, 56–80% of chronic respiratory disease cases, and 22% of cardiovascular disease (CVD) cases (3). Furthermore, it is estimated that smoking contributes to ~6 million premature mortalities worldwide each year (4).

CVDs are a significant clinical issue based upon the associated risk of mortality. In the United States, the total number of individuals affected by coronary artery disease and stroke were 15.5 and 6.6 million in 2012, respectively (5). Smoking is one of the important risk factors for the development of CVDs, and is probably the most complex and the least understood among the risk factors of CVDs (6). Cigarette smoke contains >4,000 different chemicals, ranging in size from atoms to particular matter (7), and smoking is associated with various systemic effects, including oxidative stress, systemic inflammation and endothelial dysfunction (1). However, the exact mechanism (or mechanisms) of the detrimental effects of smoking on the development of CVDs remains unknown.

Previous studies demonstrated the association of smoking with various common diseases, including hypertension, dyslipidemia, and type 2 diabetes mellitus (T2DM) (8–10). Given that certain common diseases, such as hypertension, T2DM, dyslipidemia, and metabolic syndrome (MetS), are major risk factors for CVDs, it was hypothesized that the association of smoking with CVDs may be attributable, at least in part, to its effects on these common diseases. The aim of the present study was to determine the association of smoking with the prevalence of various common diseases and metabolic parameters in community-dwelling Japanese individuals.

Subjects and methods

Study population

The study subjects comprised 5,959 community-dwelling Japanese individuals (1,302 current smokers, 1,418 past smokers and 3,239 nonsmokers) who were recruited to a population-based cohort study (Inabe Health and Longevity Study; Inabe General Hospital, Inabe, Japan), between 2010 and 2012. The Inabe Health and Longevity Study was a longitudinal genetic/epidemiological study of atherosclerotic, cardiovascular and metabolic diseases (11–14). The study protocol complied with the Declaration of Helsinki and was approved by the Ethics Committees at Mie University Graduate School of Medicine (Tsu, Japan) and Inabe General Hospital. Written informed consent was obtained from all participating subjects.

Data collection and measurements

The clinical data collected in the present study included medical history, physical examination, anthropometric measurements, and self-reported questionnaires on lifestyle (e.g., smoking habit). The section regarding smoking habit consisted of questions on smoking status, average number of cigarettes smoked per day and duration of smoking habit. Physicians or trained nurses conducted all of the following measurements: Blood pressure was measured at least twice at rest in the sitting position for >5 min; the measurements were taken by a physician or trained nurse according to the guidelines of the American Heart Association (15). Venous blood samples (10 ml) were collected in the early morning after overnight fasting. Blood samples were centrifuged at 1,600 × g for 15 min at 4°C to separate the serum. Fasting plasma glucose level, blood glycosylated hemoglobin content, and the serum concentrations of total cholesterol, triglycerides, high-density lipoprotein (HDL)-cholesterol, low-density lipoprotein (LDL) cholesterol, creatinine, and uric acid were measured as described previously (12,13).

Definitions

A smoker was defined as a subject who smoked ≥1 cigarette/day for ≥1 year. Current smoker was defined as an individual that had smoked for at least the last year. A past smoker was defined as a subject who had stopped smoking ≥1 year before enrollment in the study. A nonsmoker was defined as a subject who had never smoked. According to the daily consumption of cigarette smoking, current smokers were divided into light (1–19 cigarettes per day) and heavy (≥20 cigarettes per day) smokers. Hypertension was defined as either systolic blood pressure (SBP) ≥140 mmHg and/or diastolic BP (DBP) ≥90 mmHg, or current treatment with antihypertensive medications. T2DM was defined as either fasting plasma glucose level ≥6.93 mmol/l (126 mg/dl), blood glycosylated hemoglobin content ≥6.5%, or current use of glucose-lowering agents. Patients with T1DM were not included in the present study. Dyslipidemia was defined as either serum concentration of triglycerides ≥1.65 mmol/l (150 mg/dl), serum HDL-cholesterol concentration <1.04 mmol/l (40 mg/dl), serum LDL-cholesterol concentration ≥3.64 mmol/l (140 mg/dl), or current use of antidyslipidemic drugs. Hypertriglyceridemia was defined as either serum triglyceride concentration ≥1.65 mmol/l or current treatment of antidyslipidemic medications for hypertriglyceridemia. Hypo-HDL-cholesterolemia was defined as either serum HDL-cholesterol concentration <1.04 mmol/l. Hyper-LDL-cholesterolemia was defined as either serum LDL-cholesterol concentration ≥3.64 mmol/l or current treatment with antidyslipidemic agents for hyper-LDL-cholesterolemia. The estimated glomerular filtration rate (eGFR) was calculated using a simplified prediction equation derived from that in the Modification of Diet in Renal Disease Study (16) and proposed by the Japanese Society of Nephrology (17): eGFR (ml/min/1.73 m2) = 194 × [age (years)]−0.287 × [serum creatinine (mg/dl)]−1.094 (x 0.739 for females). Chronic kidney disease (CKD) was diagnosed with eGFR of <60 ml/min/1.73 m2, as recommended by the National Kidney Foundation Kidney Disease Outcomes Quality Initiative guidelines (18). Hyperuricemia was represented by a serum concentration of uric acid >416 µmol/l (7.0 mg/dl) or current treatment with uric acid-lowering medications. Obesity was diagnosed as body mass index (BMI) ≥25 kg/m2, based on the BMI criteria of obesity for Japanese and Asian populations (19). Diagnosis of MetS was based on the modified version of the definition proposed by the International Diabetes Federation Task Force on Epidemiology and Prevention; National Heart, Lung, and Blood Institute; American Heart Association; World Heart Federation; International Atherosclerosis Society; and International Association for the Study of Obesity (20). Cut-off values for waist circumference were used, which were ≥90 cm for men and ≥80 cm for women, based on the recommendation of the International Diabetes Association (20). A total of 1,546 subjects with MetS had three or more of the following five components: i) Waist circumference ≥90 cm for men and ≥80 cm for women; ii) serum concentration of triglycerides ≥1.65 mmol/l (150 mg/dl) or receiving drug treatment for elevated triglycerides; iii) serum concentration of HDL-cholesterol <1.04 mmol/l (40 mg/dl) for men and <1.30 mmol/l (50 mg/dl) for women; iv) SBP ≥130 mmHg or DBP of ≥85 mmHg, or drug treatment for hypertension; and v) fasting plasma glucose level ≥5.50 mmol/l (100 mg/dl) or use of glucose-lowering agents. History of obesity, dyslipidemia, hypertension or DM was evaluated by detailed questionnaire. The control subjects comprised 1,841 individuals who exhibited none of the five components of the diagnostic criteria for MetS.

Statistical analysis

Categorical variables were compared by the χ2 test. The distribution of continuous variables was examined by Kolmogorov-Smirnov-Lilliefors test. Comparison between two groups was conducted by the unpaired Student's t-test (for variables with normal distribution) or by Mann-Whitney U test (for variables with skewed distribution). Comparisons among three groups were examined by one-way analysis of variance (for variables with normal distribution) or Kruskal-Wallis test (for variables with skewed distribution). Metabolic parameters were compared between current or past smokers and nonsmokers by multivariable regression analysis following adjustment for age in men and women. Metabolic parameters were also compared between heavy and light smokers by multivariable regression analysis following adjustment for age and sex. The prevalence of common diseases was compared between current or past smokers and nonsmokers by multivariable logistic regression analysis following adjustment for age in men and women. In addition, a stepwise forward selection procedure was performed to examine the effects of smoking and age on common diseases with which smoking was associated in men and women by multivariable logistic regression analysis. The P-value for inclusion in and exclusion from the model were 0.25 and 0.1, respectively. Bonferroni's correction was applied to compensate for multiple comparisons and P<0.0009 (0.05/54) was considered to indicate a statistically significant difference. Statistical tests were conducted using the JMP software (version 5.1; SAS Institute, Inc., Cary, NC, USA).

Results

Characteristics of study population

Table I presents the characteristics of subjects according to the smoking habit. There were significant (P<0.0009) differences in age, sex, and the prevalence of hypertension, dyslipidemia, T2DM and hyperuricemia, as well as various metabolic parameters among current smokers, past smokers and nonsmokers. The number of cigarettes smoked per day was significantly lower (P=2.8×10−8) in the current smokers than in past smokers, whereas the duration of smoking habit was longer in the current smokers than in past smokers (P<1.0×10−61). With regard to the current smokers, the number of cigarettes smoked per day (19.0±8.6 vs. 12.9±5.5 cigarettes per day; P<1.0×10−61) and duration of smoking habit (28.8±11.3 vs. 19.5±9.8 years; P<1.0×10−61) were significantly greater in men than in women (Table II). Amongst past smokers, the number of cigarettes smoked per day (22.1±11.4 vs. 13.0±7.1 cigarettes per day; P<1.0×10−61) and duration of smoking habit (22.1±11.6 vs. 14.3±9.8 years; P=4.9×10−13) were also significantly greater in men than women.

Table I. Characteristics of current smokers, past smokers and nonsmokers.

Table I.

Characteristics of current smokers, past smokers and nonsmokers.

Variable	Current smokers	Past smokers	Nonsmokers	P-value
Subjects, n	1,302	1,418	3,239
Age, years	50.4±12.1	55.6±12.3	54.8±13.1	6.1×10−32
Sex, male, %	88.8	88.2	27.6	<1.0×10−61
Daily consumption, cigarettes per day	18.3±8.5	20.1±11.3		2.8×10−8
Duration of smoking habit, years	27.7±11.5	21.2±11.6		<1.0×10−61
Hypertension, %	31.0	44.9	35.3	3.1×10−14
Dyslipidemia, %	68.0	68.5	58.2	1.4×10−14
Type 2 diabetes mellitus, %	13.2	15.3	9.4	5.2×10−9
Hyperuricemia, %	25.4	28.9	8.6	<1.0×10−61
Body weight, kg	66.3±12.3	65.7±11.3	56.7±10.7	<1.0×10−61
Body mass index, kg/m2	23.4±3.6	23.6±3.2	22.5±3.4	9.4×10−33
Waist circumference, cm	82.2±9.2	82.6±8.6	78.7±9.2	1.0×10−53
Systolic blood pressure, mmHg	118.5±15.4	123.1±15.8	120.1±16.0	1.1×10−14
Diastolic blood pressure, mmHg	74.8±12.4	77.9±11.7	73.0±11.8	3.8×10−37
Serum triglycerides, mmol/l	1.54±1.12	1.41±0.98	1.09±0.61	<1.0×10−61
Serum HDL-cholesterol, mmol/l	1.47±0.39	1.59±0.43	1.77±0.44	<1.0×10−61
Serum LDL-cholesterol, mmol/l	3.15±0.86	3.19±0.78	3.18±0.78	0.2834
Fasting plasma glucose, mmol/l	5.69±1.37	5.77±1.24	5.41±0.86	1.8×10−46
Blood glycosylated hemoglobin, %	5.76±0.87	5.71±0.73	5.65±0.54	0.1109
Serum uric acid, µmol/l	356.6±82.0	360.7±82.4	295.3±75.5	<1.0×10−61
Serum creatinine, µmol/l	71.0±13.7	74.8±19.1	61.6±14.9	<1.0×10−61
eGFR, ml/min/1.73 m−2	80.6±14.9	74.7±14.7	77.2±15.4	3.9×10−23

[i] Data are presented as means ± standard deviation. Categorical variables were compared using the χ² test; continuous variables were compared using the Mann-Whitney U test between two groups and by the Kruskal-Wallis test among three groups (due to skewed distribution); P<0.0009 was considered to indicate a statistically significant difference. eGFR = 194 × [age (years)]^−0.287 × [serum creatinine (mg/dl)]^−1.094 (x 0.739 for female). HDL, high-density lipoprotein; LDL, low-density lipoprotein; eGFR, estimated glomerular filtration rate.

Table II. Comparisons of metabolic parameters among current smokers, past smokers, and nonsmokers in men and women.

Table II.

Comparisons of metabolic parameters among current smokers, past smokers, and nonsmokers in men and women.

A, Men
Variable	Current smokers	Past smokers	Nonsmokers	P-valuea	P-valueb
Subjects, n	1,156	1,251	894
Age, years	50.9±12.0	56.6±12.2	54.3±15.2
Daily consumption, cigarettes per day	19.0±8.6c	22.1±11.4d
Duration of smoking habit, years	28.8±11.3e	22.1±11.6f
Body weight, kg	67.5±11.9	67.4±10.7	65.8±10.9	0.1077	2.7×10−6
Body mass index, kg/m2	23.4±3.5	23.8±3.2	23.4±3.3	0.8567	0.0006
Waist circumference, cm	82.7±9.0	83.5±8.3	81.9±8.7	0.0477	0.0002
Systolic blood pressure, mmHg	119.1±15.5	124.5±15.4	122.9±15.7	4.6×10−5	0.1568
Diastolic blood pressure, mmHg	75.4±12.4	78.9±11.4	76.9±12.4	0.0647	0.0003
Serum triglycerides, mmol/l	1.59±1.16	1.46±1.01	1.24±0.73	7.8×10−15	5.7×10−8
Serum HDL-cholesterol, mmol/l	1.43±0.37	1.56±0.42	1.55±0.39	1.3×10−11	0.4316
Serum LDL-cholesterol, mmol/l	3.18±0.86	3.21±0.77	3.18±0.77	0.6802	0.2247
Fasting plasma glucose, mmol/l	5.75±1.40	5.83±1.27	5.65±1.00	0.0180	0.0021
Blood glycosylated hemoglobin, %	5.78±0.90	5.74±0.74	5.70±0.63	0.0014	0.4516
Serum uric acid, µmol/l	366.2±78.4	371.8±78.3	359.0±76.5	0.0955	0.0002
Serum creatinine, µmol/l	73.0±12.6	77.2±18.8	76.9±14.2	2.6×10−7	0.8858
eGFR, ml/min/1.73 m−2	80.1±14.7	74.0±14.6	74.7±15.6	1.5×10−6	0.8169
B, Women
Variable	Current smokers	Past smokers	Nonsmokers	P-valuea	P-valueb
Subjects, n	146	167	2,345
Age, years	45.8±11.8	48.2±11.1	55.0±12.2
Daily consumption, cigarettes per day	12.9±5.5c	13.0±7.1d
Duration of smoking habit, years	19.5±9.8e	14.3±9.3f
Body weight, kg	56.9±11.5	53.4±8.3	53.3±8.4	0.0003	0.4382
Body mass index, kg/m2	22.7±4.3	21.7±3.0	22.2±3.3	0.0046	0.3784
Waist circumference, cm	78.6±9.8	76.4±8.4	77.6±9.1	0.0049	0.6784
Systolic blood pressure, mmHg	113.7±13.8	113.0±14.7	119.1±16.0	0.7554	0.0428
Diastolic blood pressure, mmHg	69.9±11.4	70.0±11.4	71.5±11.3	0.8556	0.6562
Serum triglycerides, mmol/l	1.09±0.58	1.02±0.62	1.03±0.55	0.0003	0.1072
Serum HDL-cholesterol, mmol/l	1.74±0.43	1.85±0.42	1.85±0.44	7.9×10−5	0.4955
Serum LDL-cholesterol, mmol/l	2.99±0.83	3.00±0.78	3.18±0.79	0.4680	0.2080
Fasting plasma glucose, mmol/l	5.30±1.06	5.32±0.84	5.33±0.78	0.1543	0.1400
Blood glycosylated hemoglobin, %	5.57±0.51	5.56±0.65	5.63±0.51	0.2508	0.5529
Serum uric acid, µmol/l	278.4±68.2	273.9±58.9	271.2±59.4	0.0038	0.0385
Serum creatinine, µmol/l	54.0±9.9	55.9±7.5	55.8±10.5	0.4956	0.2285
eGFR, ml/min/1.73 m−2	85.0±16.4	80.0±14.0	78.1±15.2	0.1232	0.0658

{ label (or @symbol) needed for fn[@id='tfn2-br-0-0-991'] } Data are presented as means ± standard deviation. Categorical variables were compared using the χ² test; multivariable regression analysis was performed after adjustment for age; P<0.0009 was considered to indicate a statistically significant difference.

a Comparison between current smokers and nonsmokers

b comparison between past smokers and nonsmokers. Daily consumption of cigarettes and duration of smoking habit were compared between men and women in current or past smokers by Mann-Whitney U test (due to variables with skewed distribution)

c P<1.0×10⁻⁶¹

d P<1.0×10⁻⁶¹

e P<1.0×10⁻⁶¹

f P=4.9×10⁻¹³. eGFR (m/min/1.73 m²) = 194 × [age (years)]^−0.287 × [serum creatinine (mg/dl)]^−1.094 (x 0.739 for female). HDL, high-density lipoprotein; LDL, low-density lipoprotein; eGFR, estimated glomerular filtration rate.

Smoking and metabolic parameters

Multivariable regression analysis after adjustment for age demonstrated significantly higher serum concentration of triglycerides (P=7.8×10−15) and eGFR (P=1.5×10−6) in current smokers than nonsmokers among men, whereas SBP (P=4.6×10−5), serum HDL-cholesterol concentration (P=1.3×10−11), and serum creatinine concentration (P=2.6×10−7) were lower in current smokers than in nonsmokers among men (Table II). In women, current smokers also had significantly higher body weights (P=0.0003) and serum triglyceride concentrations (P=0.0003), as well as lower serum HDL-cholesterol concentration (P=7.9×10−5), compared with nonsmokers. In men, past smokers had significantly higher body weights (P=2.7×10−6), BMI (P=0.0006), waist circumference (P=0.0002), and DBP (P=0.0003), as well as higher serum concentrations of triglycerides (P=5.7×10−8) and uric acid (P=0.0002), compared with nonsmokers.

Subsequently, a battery of metabolic parameters was compared following adjustment for age between smokers and nonsmokers who did not take any medications for hypertension, dyslipidemia, T2DM or hyperuricemia. In men who were not taking antihypertensive agents, SBP was significantly lower in current smokers than nonsmokers (P=3.1×10−5; Table III). In men not taking any lipid-lowering medications, serum triglyceride concentrations were significantly higher (P=3.3×10−14) and serum HDL-cholesterol concentration was significantly lower (P=6.3×10−9) in current smokers than nonsmokers, whereas past smokers exhibited higher serum triglyceride concentrations compared with nonsmokers (P=3.1×10−8). Fasting plasma glucose level (P=0.0009) and serum uric acid concentration (P=0.0007) were also higher in past smokers than nonsmokers in men not taking any medications for T2DM or hyperuricemia. In women not taking any lipid-lowering medications, serum triglyceride concentrations (P=8.3×10−6) and serum HDL-cholesterol concentrations (P=6.5×10−5) were significantly higher and lower, respectively, in current smokers than nonsmokers.

Table III. Comparisons of metabolic parameters between (current or past) smokers and nonsmokers among men and women not on medications for hypertension, dyslipidemia, type 2 diabetes mellitus or hyperuricemia.

Table III.

Comparisons of metabolic parameters between (current or past) smokers and nonsmokers among men and women not on medications for hypertension, dyslipidemia, type 2 diabetes mellitus or hyperuricemia.

A, Men
Metabolic parameters	Current smokers	Past smokers	Nonsmokers	P-valuea	P-valueb
Individuals not on antihypertensive medications
Subjects, n	946	861	665
Systolic blood pressure, mmHg	117.1±15.0	121.5±14.8	120.6±15.3	3.1×10−5	0.8965
Diastolic blood pressure, mmHg	74.1±12.0	77.6±11.3	75.9±12.1	0.0083	0.0248
Individuals not on antidyslipidemic medications
Subjects, n	1,019	1,016	725
Serum triglycerides, mmol/l	1.55±1.13	1.43±1.04	1.19±0.68	3.3×10−14	3.1×10−8
Serum HDL-cholesterol, mmol/l	1.45±0.38	1.58±0.43	1.56±0.39	6.3×10−9	0.2799
Serum LDL-cholesterol, mmol/l	3.18±0.86	3.26±0.77	3.21±0.75	0.4909	0.1459
Individuals not on antidiabetic medications
Subjects, n	1,067	1,127	805
Fasting plasma glucose, mmol/l	5.56±0.90	5.64±1.00	5.50±0.60	0.0306	0.0009
Subjects, n	636	831	608
Blood glycosylated hemoglobin, %	5.62±0.50	5.62±0.62	5.60±0.43	0.0227	0.5270
Individuals not on uric acid-lowering medications
Subjects, n	1,018	1,136	761
Serum uric acid, µmol/l	364.8±76.7	370.5±77.6	358.7±76.2	0.1638	0.0007
B, Women
Metabolic parameters	Current smokers	Past smokers	Nonsmokers	P-valuea	P-valueb
Individuals not on antihypertensive medications
Subjects, n	130	149	1,868
Systolic blood pressure, mmHg	112.2±13.2	111.0±13.4	116.1±15.0	0.8657	0.0265
Diastolic blood pressure, mmHg	69.1±11.4	68.7±10.8	70.1±11.0	0.8412	0.4739
Individuals not on antidyslipidemic medications
Subjects, n	135	150	1,870
Serum triglycerides, mmol/l	1.09±0.59	0.96±0.61	0.97±0.51	8.3×10−6	0.1883
Serum HDL-cholesterol, mmol/l	1.74±0.43	1.89±0.42	1.88±0.44	6.5×10−5	0.8397
Serum LDL-cholesterol, mmol/l	2.99±0.82	2.96±0.78	3.21±0.80	0.5661	0.0602
Individuals not on antidiabetic medications
Subjects, n	144	159	2,238
Fasting plasma glucose, mmol/l	5.27±1.04	5.21±0.53	5.25±0.61	0.0467	0.5654
Subjects, n	84	121	1,862
Blood glycosylated hemoglobin, %	5.52±0.39	5.45±0.37	5.58±0.43	0.3485	0.2502
Individuals not on uric acid-lowering medications
Subjects, n	130	153	2,109
Serum uric acid, µmol/l	278.4±68.2	273.1±58.2	271.0±59.3	0.0034	0.0556

{ label (or @symbol) needed for fn[@id='tfn9-br-0-0-991'] } Data are presented as means ± standard deviation. Multivariable regression analysis was performed after adjustment for age; P<0.0009 was considered to indicate a statistically significant difference.

a Comparison between current smokers and nonsmokers

b comparison between past smokers and nonsmokers. HDL, high-density lipoprotein; LDL, low-density lipoprotein.

Smoking and the prevalence of common diseases

Multivariable logistic regression analysis following adjustment for age identified current smoking to be significantly associated with the prevalence of dyslipidemia [P=6.3×10−10; odds ratio (OR), 1.81], hypertriglyceridemia (P=2.3×10−20; OR, 2.39), hypo-HDL-cholesterolemia (P=2.0×10−9; OR, 2.14), MetS (P=0.0003; OR 1.61), and CKD (P=4.4×10−5; OR, 0.54) in men (Table IV). Past smoking was significantly associated with dyslipidemia (P=3.6×10−8; OR, 1.68), hypertriglyceridemia (P=2.3×10−11; OR, 1.85), hyper-LDL-cholesterolemia (P=5.5×10−5; OR, 1.43), MetS (P=4.2×10−5; OR, 1.71), and hyperuricemia (P=0.0002; OR, 1.37) in men. In women, neither current nor past smoking was associated with any common diseases.

Table IV. Results of multivariable logistic regression analysis of smoking status and prevalence of common diseases in men and women.

Table IV.

Results of multivariable logistic regression analysis of smoking status and prevalence of common diseases in men and women.

A, Men
	Current smokers		Past smokers
Common disease	Odds ratio (95% confidence interval)	P-value	Odds ratio (95% confidence interval)	P-value
Hypertension	0.83 (0.68–1.01)	0.0650	1.19 (0.99–1.44)	0.0685
Type 2 diabetes mellitus	1.33 (1.02–1.73)	0.0381	1.22 (0.95–1.57)	0.1184
Dyslipidemia	1.81 (1.50–2.19)	6.3×10−10	1.68 (1.39–2.01)	3.6×10−8
Hypertriglyceridemia	2.39 (1.99–2.88)	2.3×10−20	1.85 (1.54–2.21)	2.3×10−11
Hypo-HDL-cholesterolemia	2.14 (1.68–2.76)	2.0×10−9	1.40 (1.09–1.81)	0.0086
Hyper-LDL-cholesterolemia	1.23 (1.03–1.46)	0.0251	1.43 (1.20–1.70)	5.5×10−5
Obesity	1.11 (0.92–1.34)	0.2695	1.33 (1.11–1.60)	0.0019
Metabolic syndrome	1.61 (1.25–2.09)	0.0003	1.71 (1.32–2.21)	4.2×10−5
Chronic kidney disease	0.54 (0.40–0.72)	4.4×10−5	0.93 (0.73–1.18)	0.5476
Hyperuricemia	1.14 (0.93–1.41)	0.2173	1.37 (1.12–1.67)	0.0002
B, Women
	Current smokers		Past smokers
Common disease	Odds ratio (95% confidence interval)	P-value	Odds ratio (95% confidence interval)	P-value
Hypertension	1.05 (0.66–1.65)	0.8230	0.64 (0.40–1.00)	0.0556
Type 2 diabetes mellitus	1.24 (0.54–2.50)	0.5718	1.36 (0.67–2.49)	0.3592
Dyslipidemia	1.05 (0.72–1.53)	0.8052	0.85 (0.59–1.20)	0.3540
Hypertriglyceridemia	1.56 (1.00–2.38)	0.0445	1.18 (0.76–1.78)	0.4359
Hypo-HDL-cholesterolemia	1.57 (0.59–3.45)	0.3066	0.83 (0.25–2.05)	0.7259
Hyper-LDL-cholesterolemia	0.75 (0.51–1.10)	0.1453	0.88 (0.62–1.24)	0.4748
Obesity	1.77 (1.21–2.56)	0.0030	0.96 (0.64–1.42)	0.8532
Metabolic syndrome	1.90 (1.10–3.20)	0.0180	0.65 (0.36–1.12)	0.1292
Chronic kidney disease	0.87 (0.37–1.86)	0.7398	0.76 (0.33–1.52)	0.4781
Hyperuricemia	2.24 (0.65–5.84)	0.1395	2.30 (0.77–5.56)	0.0904

[i] Multivariable logistic regression analysis was performed after adjustment for age; P<0.0009 was considered to indicate a statistically significant difference. HDL, high-density lipoprotein; LDL, low-density lipoprotein.

Subsequently, a stepwise forward selection procedure was performed to examine the effects of smoking and age on various common diseases that were found to be associated with smoking by multivariable logistic regression analysis. The analysis found current smoking to be a significant (P<0.0009) and independent determinant of dyslipidemia, hypertriglyceridemia, hypo-HDL-cholesterolemia, MetS, and CKD in men (Table V). Past smoking was also a significant (P<0.0009) and independent determinant of dyslipidemia, hypertriglyceridemia, hyper-LDL-cholesterolemia and MetS in men.

Table V. Effects of smoking on various metabolic disorders in men determined by a stepwise forward selection procedure.

Table V.

Effects of smoking on various metabolic disorders in men determined by a stepwise forward selection procedure.

A, Current smokers vs. nonsmokers
Disease	Variable	R2	P-value
Dyslipidemia	Age	0.0121	3.8×10−10
	Smoking	0.0147	6.3×10−10
Hypertriglyceridemia	Age	0.0292	0.0007
	Smoking	0.0041	2.3×10−20
Hypo-HDL-cholesterolemia	Age	0.0183	0.0117
	Smoking	0.0034	2.0×10−9
Metabolic syndrome	Age	0.0534	1.4×10−18
	Smoking	0.0089	0.0003
Chronic kidney disease	Age	0.2096	3.3×10−45
	Smoking	0.0107	4.4×10−5
B, Past smokers vs. nonsmokers
Disease	Variable	R2	P-value
Dyslipidemia	Age	0.0112	3.9×10−8
	Smoking	0.0132	3.6×10−8
Hypertriglyceridemia	Age	0.0016	0.0312
	Smoking	0.0166	2.3×10−11
Hyper-LDL-cholesterolemia	Age	NA	0.4028
	Smoking	0.0058	3.9×10−5
Metabolic syndrome	Age	0.0659	3.6×10−19
	Smoking	0.0109	4.2×10−5
Hyperuricemia	Age	0.0011	0.1046
	Smoking	0.0038	0.0020

[i] P<0.0009 was considered to indicate a statistically significant difference. R², contribution rate; NA, not applicable; HDL, high-density lipoprotein; LDL, low-density lipoprotein.

Comparisons between heavy and light smokers

The differences in various metabolic parameters were compared between the light and heavy current smokers. The percentage of males (P=2.0×10−15), number of cigarettes smoked per day (P<1.0×10−61), and duration of smoking habit (P=2.0×10−10) were significantly higher in heavy smokers than light smokers (Table VI). Analysis following adjustment for age and sex demonstrated that body weight (P=0.0005), BMI (P=3.7×10−6), waist circumference (P=5.3×10−6), and serum concentration of triglycerides (P=0.0009) were significantly higher in heavy smokers than in light smokers, whereas serum concentrations of HDL-cholesterol were significantly lower in heavy smokers than in light nonsmokers (P=4.4×10−5).

Table VI. Comparison of metabolic parameters between light and heavy smokers.

Table VI.

Comparison of metabolic parameters between light and heavy smokers.

Variable	Light smokers	Heavy smokers	P-valuea	P-valueb
Subjects, n	523	721
Age, years	48.8±12.1	50.3±11.2	0.0111	–
Sex, male, %	80.5	94.9	2.0×10−15	–
Daily consumption, cigarettes per day	11.0±3.7	23.6±7.1	<1.0×10−61	–
Duration of smoking habit, years	25.2±12.2	29.5±10.7	2.0×10−10	–
Body weight, kg	64.6±11.8	68.2±12.3	6.3×10−7	0.0005
Body mass index, kg/m2	22.9±3.5	23.9±3.7	1.7×10−7	3.7×10−6
Waist circumference, cm	80.6±8.8	83.5±9.3	7.8×10−9	5.3×10−6
Systolic blood pressure, mmHg	117.1±15.2	119.3±15.5	0.0067	0.1788
Diastolic blood pressure, mmHg	73.8±12.4	75.7±12.4	0.0107	0.1612
Serum triglycerides, mmol/l	1.38±0.99	1.67±1.22	1.5×10−9	0.0009
Serum HDL-cholesterol, mmol/l	1.54±0.40	1.41±0.37	2.2×10−9	4.4×10−5
Serum LDL-cholesterol, mmol/l	3.13±0.88	3.17±0.83	0.1935	0.8931
Fasting plasma glucose, mmol/l	5.57±1.37	5.79±1.41	1.1×10−5	0.0736
Blood glycosylated hemoglobin, %	5.67±0.85	5.84±0.92	4.3×10−5	0.0899
Serum uric acid, µmol/l	348.4±86.5	361.9±78.9	0.0070	0.8480
Serum creatinine, µmol/l	69.5±13.6	71.8±12.4	0.0032	0.4383
eGFR, ml/min/1.73 m−2	81.2±15.1	80.6±14.4	0.6302	0.5820

{ label (or @symbol) needed for fn[@id='tfn14-br-0-0-991'] } Data are presented as means ± standard deviation.

a Categorical variables were compared by the χ² test; continuous variables were compared by the Mann-Whitney U test (due to skewed distribution)

b multivariable regression analysis was performed after adjustment for age and sex; P<0.0009 was considered to indicate a statistically significant difference. eGFR (m/min/1.73 m²) = 194 × [age (years)]^−0.287 × [serum creatinine (mg/dl)]^−1.094 (x 0.739 for female). HDL, high-density lipoprotein; LDL, low-density lipoprotein; eGFR, estimated glomerular filtration rate.

Discussion

In the present study, the association of smoking with the prevalence of various common diseases and metabolic parameters was examined in community-dwelling Japanese individuals. This cross-sectional study demonstrated that smoking was significantly associated with higher serum triglyceride concentrations and lower HDL-cholesterol concentrations in men and women. Additionally, smoking correlated significantly with lower serum creatinine concentrations and SBP, as well as higher eGFR in men and with higher body weight in women. The current results demonstrated that smoking is significantly associated with the prevalence of dyslipidemia, hypertriglyceridemia, hypo-HDL-cholesterolemia, MetS, and CKD in men, although not in women. Furthermore, heavy smoking was associated with significantly greater serum triglyceride concentrations and reduced serum HDL-cholesterol concentrations, as well as increased body weight, BMI, and waist circumference compared with light smokers.

The association between smoking and serum lipids has been reported in numerous studies since the 1980s (9,21,22). These studies demonstrated that smoking decreased serum HDL-cholesterol levels and increased serum triglyceride and LDL-cholesterol levels. Furthermore, passive smoking was demonstrated to be associated with low HDL-cholesterol levels in children (23). A meta-analysis of cross-sectional studies indicated that smokers had increased serum concentrations of total cholesterol (3.0%), triglycerides (9.1%), very-LDL cholesterol (10.4%), and LDL-cholesterol (1.7%) and lower serum concentrations of HDL-cholesterol (−5.7%) and apolipoprotein A1 (−4.2%), compared with nonsmokers (9). The studies also demonstrated a dose-response association between the quantity of cigarette smoking and the extent of lipid abnormalities. The present findings are consistent with the above-mentioned previous reports. In addition, the current study indicated that past smoking is associated with dyslipidemia and hypertriglyceridemia in men. Cessation of smoking is reported to have relatively limited effects on lipid profiles (24,25). A meta-analysis showed that smoking cessation led to increased serum concentration of HDL-cholesterol, but no significant improvement in serum triglycerides and LDL-cholesterol (25).

The mechanisms by which cigarette smoking alters serum lipid profiles are not fully understood. One popular theory is that nicotine stimulates the secretion of catecholamines, as well as other hormones (cortisol and growth hormones), leading to increased serum concentrations of free fatty acids, which stimulate hepatic secretion of very LDL and triglycerides (26). Other proposed explanations include smoking-induced increases in cholesteryl ester transfer protein, reduction of lectin cholesterol acyltransferase activity, and alteration of apolipoprotein A1 synthesis, resulting in changes in serum lipid profiles (27). As dyslipidemia, including hypertriglyceridemia and hypo-HDL-cholesterolemia, is important in the pathogenesis of atherosclerosis, smoking may accelerate the development of CVDs indirectly via its detrimental effects on serum lipid profiles. Smoking was also reported to induce qualitative changes in serum lipids (28–30). Excess free radicals and oxidants present in cigarette smoke, as well as endogenously produced oxidants and radicals, enhance the pro-oxidative environment (6,29). It was suggested that peroxynitrite is significant in the oxidative modification of plasma LDL induced by smoking (30). The oxidation of lipids may be another way by which smoking accelerates the development of CVDs.

The current results demonstrated that smoking was significantly associated with reduced SBP in men. However, there are conflicting data regarding the association between smoking and BP. Various studies in the 1970s reported no significant differences in BP between smokers and nonsmokers (31,32), whereas a more recent study demonstrated that BP was significantly increased in elderly male smokers (8). Furthermore, a recent meta-analysis indicated that smoking was associated with lower SBP and DBP, and reduced risk of hypertension in observational analyses, although observational and Mendelian randomization analyses did not support a causal association between heaviness of smoking and BP among current smokers (33). Thus, the association between smoking and hypertension remains elusive.

MetS is a complex disease involving abdominal obesity, impaired glucose tolerance, hypertension and dyslipidemia (20). The associations between MetS and increased risk of CVDs, morbidity, and mortality have been established (34,35). The current results demonstrated that smoking was significantly associated with the prevalence of MetS in Japanese men. In addition, previous studies demonstrated a significant association between smoking and MetS (36,37). Smoking increases the risk of dyslipidemia (9,21,22) and causes glucose intolerance (38), leading to an increased risk of T2DM (10). Furthermore, smoking is associated with central fat accumulation (39,40). These observations indicate that smoking affects multiple components of the diagnostic criteria for MetS and enhances the development of CVDs.

CKD is a risk factor for end-stage renal disease, as well as for CVDs (41). The present study demonstrated that smoking was significantly associated with higher eGFR and lower prevalence of CKD in men. Previous cross-sectional studies also reported that smokers exhibited elevated GFRs compared with nonsmokers (42,43). Smoking causes glucose intolerance (38), whereas higher insulin resistance was demonstrated to be associated with increased GFR (44). These observations indicate that smoking induces a rise in GFR, at least in part, via its effects on glucose intolerance. However, recent longitudinal studies have indicated that smoking is associated with increased long-term risks of reduced GFR, glomerular hyperfiltration and proteinuria (45,46). As smoking increases the risk of T2DM and obesity, it may induce diabetic nephropathy or obesity-associated glomerulopathy, resulting in impaired renal function in the long-term. Other proposed explanations include smoking-induced renal atherosclerosis (47), impaired endothelial function (48), and alterations in systemic and renal hemodynamics (49), resulting in chronic kidney damage.

The current results demonstrated certain differences in the association of smoking with metabolic parameters, and the prevalence of common diseases between men and women. These findings may be associated with differences in the quantity of smoking between the two sexes. Notably, numerous previous studies revealed sex differences in the effect of smoking on metabolic parameters (50–52). One longitudinal study of non-diabetic individuals demonstrated that smoking was associated with improved insulin sensitivity in women, but reduced insulin sensitivity in men (50). Reed et al (51) reported the association of passive smoking with greater BMIs and fasting plasma glucose levels in men, and with reduced HDL-cholesterol levels in women among the Old Order Amish population. In late adolescence, passive smoking was also associated with lower HDL-cholesterol in girls, demonstrating a sex difference in the findings (52). Furthermore, Ahonen et al (53) reported the different association of smoking with subclinical inflammation between sexes, with a decreased adiponectin level in women, but increased high-sensitivity C-reactive protein levels in men. Smoking is reported to be associated with lower estrogen concentrations in women (54), whereas estrogen deficiency promotes metabolic dysfunction, and predisposes individuals to obesity, MetS and T2DM (55). The effects of smoking on sex hormones may be one explanation by which smoking affects metabolic parameters with sex differences. However, the molecular mechanism underlying the observed sex differences in the effects of smoking on metabolic parameters has not been elucidated.

There were various limitations of the present study: i) As the results of the present study were not replicated, validation of the findings is required by other independent subject panels; ii) based upon the cross-sectional study design, causalities could not be confirmed; iii) data regarding other daily habits, including diet, alcohol consumption and exercise, were not evaluated and these certainly affect various metabolic parameters; iv) given that the percentages of current and past smokers were low in women (current smokers, 5.5%; past smokers, 6.3%), the statistical power may not be enough to estimate certain metabolic parameters or common diseases.

In conclusion, the current cross-sectional study demonstrated that smoking was significantly associated with various metabolic abnormalities and the prevalence of common diseases in Japanese individuals, with certain sex differences. These observations indicate that smoking affects multiple metabolic parameters, and potentially accelerates the development of CVDs with certain sex differences.

Acknowledgements

The present study was supported by CREST (grant no. JPMJCR1302) of the Japan Science and Technology Agency, and by the Japan Society for the Promotion of Science KAKENHI (grant no. JP15H04772).

References