Relationship Between Smoking and Metabolic Syndrome in Men with Diabetes Mellitus

Abstract

Introduction:

Metabolic syndrome is a potent risk factor for cardiovascular disease and is often complicated in patients with diabetes. The aim of this study was to determine whether and how smoking is related to metabolic syndrome in patients with diabetes.

Methods:

Subjects were men with diabetes (n=2675) who had been diagnosed as a result of health examinations at their workplaces. They were divided into nonsmokers, light smokers (≤20 cigarettes per day) and heavy smokers (>20 cigarettes per day). The relationships of smoking with metabolic syndrome and its components, including obesity, high blood pressure, and dyslipidemia [high triglycerides and/or low high-density lipoprotein cholesterol (HDL-C)] in nondrinkers, light drinkers (<22 grams of ethanol per day), and heavy drinkers (≥22 grams of ethanol per day) were investigated by logistic regression analysis with adjustment for age, regular exercise, and drug therapy for diabetes.

Results:

In nondrinkers with diabetes, the odds ratio (OR) versus nonsmokers for metabolic syndrome was significantly higher in heavy smokers (2.47 [95% confidence interval (CI) 1.43–4.25]) but not different in light smokers [1.03 (95% CI 0.71–1.49)] when compared with the reference level of 1.00. In the light and heavy drinker groups, the ORs versus nonsmokers for metabolic syndrome were not significantly different from the reference level in the smoker groups, except for a significantly lower OR of light smokers in the heavy drinker group [0.69 (95% CI 0.53–0.90)]. In nondrinkers, the ORs of heavy smokers versus nonsmokers for large waist circumference [1.64 (95% CI 1.01–2.68)], high triglycerides [1.78 (95% CI 1.12–2.81)], and low HDL-C [3.20 (95% CI 1.99–5.14]), but not the OR for high blood pressure [1.02 (95% CI 0.60–1.72)], were significantly higher than the reference level.

Conclusion:

In nondrinkers with diabetes, there is a positive association between heavy smoking and metabolic syndrome, which is mainly due to higher risks of central obesity and dyslipidemia, such as hypertriglyceridemia and low HDL cholesterolemia, in heavy smokers than in nonsmokers.

Introduction

Metabolic syndrome is a cluster of risk factors, including insulin resistance, central obesity, dyslipidemia, and hypertension, for the development of cardiovascular disease (CVD). Reduction of risk factors including smoking is important for prevention of CVD. In most previous studies on the relationship between smoking and metabolic syndrome, a positive association between them has been shown,^1
–3 whereas one study showed an inverse association in Turkish women.⁴ A recent meta-analysis study using data from 13 prospective studies indicated that the relative risk of active smoking for metabolic syndrome was 1.26 (1.10–1.44) with a statistical significance.⁵ The higher risk of metabolic syndrome in smokers may be mainly explained by proneness to dyslipidemia, such as higher triglycerides and lower high-density lipoprotein cholesterol (HDL-C), in smokers.⁶ The major cause of death for patients with diabetes is CVD, of which the risk is greatly increased by complication of metabolic syndrome.^7
–9 The incidences of CVD have been reported to be comparable in diabetic and nondiabetic persons not complicated with metabolic syndrome.¹⁰ However, it remains unknown whether and how the risk of metabolic syndrome is modified by smoking in patients with diabetes.

Therefore, the purpose of this study was to clarify the relationship between smoking and metabolic syndrome in patients with diabetes. Two international sets of criteria have been generally used for diagnosis of metabolic syndrome. One set is the criteria proposed by the National Cholesterol Education Program Adult Treatment Panel III (NCEP ATP III)¹¹ and the other set is that by the International Diabetes Federation (IDF).¹² In the present study, relationships of smoking with metabolic syndrome diagnosed by the above two criteria were investigated separately. Light-to-moderate alcohol drinkers are known to have a lower risk than nondrinkers of developing CVD,¹³ which is in part explained by higher HDL-C levels in drinkers than in nondrinkers.¹⁴ Although there is still debate regarding the relation of alcohol to metabolic syndrome, according to a recent meta-analysis study, the risk of metabolic syndrome in general populations was lower in light-to-moderate alcohol drinkers (less than 40 grams of ethanol/day) than in nondrinkers with the odds ratio (OR) for metabolic syndrome being 0.84 (0.75–0.94) in men.¹⁵ On the other hand, in patients with diabetes, heavy drinking (44 grams of ethanol/day or more) has been reported to be associated with increased risk of metabolic syndrome.¹⁶ Thus, there is a possibility that the relationship of smoking with metabolic syndrome is modified by alcohol drinking as well as diabetes. Accordingly, in the present study, the relationship of smoking with metabolic syndrome was analyzed separately in nondrinkers, light drinkers, and heavy drinkers.

Methods

Subjects

A cross-sectional study was performed using a local population-based database. The subjects in the original database of the health checkup were male workers aged from 35 to 70 years (n=37,693) who had received periodic health examinations at their workplaces in Yamagata Prefecture in Japan. All of the subjects were of Japanese origin. Subjects who were receiving treatment for any illness were requested to state the names of the diseases in a questionnaire at the health checkup. Subjects with diabetes (n=2675) were extracted from the database according to the definition of diabetes given below. Subjects with diabetes were defined as those showing high glycated hemoglobin (HbA1c levels) (≥6.5%), according to the recent criteria for diagnosis of diabetes by the American Diabetes Association,¹⁷ and/or having a current history of drug therapy for diabetes. This study was approved by the Ethics Committee of Yamagata University School of Medicine. Histories of cigarette smoking, alcohol consumption, regular exercise (almost every day with exercise for 30 min or longer per day), and illness were also surveyed by questionnaires.

In the self-written questionnaire paper, subjects were first asked, “Are you a habitual cigarette smoker?” Cigarette smokers were defined as subjects who had smoked for 6 months or longer and had smoked for the past month or longer. Then the subjects who were smokers were further asked, “What is your average cigarette consumption per day?” The response categories for this question were “20 cigarettes or less per day,” “more than 20 and less than 41 cigarettes per day,” and “41 or more cigarettes per day.” In this study, the subjects were divided into three groups by average cigarette consumption (nonsmokers; light smokers, ≤20 cigarettes per day; heavy smokers, >20 cigarettes per day).

Usual daily alcohol consumption was calculated in terms of the equivalent number of “go,” a traditional Japanese unit of amount of sake (rice wine). The amounts of other alcoholic beverages, including beer, wine, whiskey, and shochu (traditional Japanese distilled spirits), were converted and expressed as units of go. One go approximately corresponds to 180 mL of sake (rice wine), 500 mL of beer, 240 mL of wine, 60 mL of whisky, and 110 mL of shochu. In the questionnaire paper, the above conversion from amount of each beverage to go was explained before the following question on the amount of alcohol consumption. In the questionnaire, subjects were asked, “What is your average alcohol consumption per drinking day?” The response categories for this question were “null,” “less than 1 go per day,” “1 go or more and less than 2 go per day,” “2 go or more and less than 3 go per day,” and “3 go or more per day.” One go contains about 22 grams of ethanol, and this amount was used to separate heavy drinkers from light drinkers because it is generally accepted that alcohol intake should be reduced to less than 20–30 grams ethanol per day from the viewpoint of prevention of hypertension.^18,19 Thus, the subjects were divided into three groups according to usual ethanol consumption per day (nondrinkers; light drinkers, <22 grams of ethanol per day; heavy drinkers, ≥22 grams ethanol per day).

Measurements

Height and body weight were measured with the subjects wearing light clothes at the health checkup. Body mass index (BMI) was calculated as weight in kilograms divided by the square of height in meters. Waist circumference (WC) was measured at the navel level according to the recommendation of the definition of the Japanese Committee for the Diagnostic Criteria of Metabolic Syndrome.²⁰ Blood pressure was measured by trained nurses, who were part of the local health checkup company, with a mercury sphygmomanometer once on the day of the health checkup after each subject had rested quietly in a sitting position. Korotkoff phase V was used to define diastolic blood pressure (DBP). Fasted blood was sampled from each subject in the morning, and serum triglycerides (TGs) and HDL-C levels were measured by enzymatic methods using commercial kits, Pureauto S TG-N and Cholestest N-HDL (Sekisui Medical Co., Ltd, Tokyo, Japan), respectively. The coefficients of variation for the reproducibility of measurement were ≤3% for TGs and ≤5% for HDL-C. HbA1c was measured by the National Glycohemoglobin Standardization Program (NGSP)-approved technique using the latex cohesion method with a commercial kit (Determiner HbA1c, Kyowa Medex, Tokyo, Japan). The coefficient of variation for reproducibility of HbA1c measurement was ≤5%. Because the standards of HbA1c used for measurement are different in the NGSP method and the Japan Diabetes Society (JDS) method, HbA1c values were calibrated by using a formula proposed by the JDS: HbA1c (NGSP) (%)=1.02×HbA1c (JDS) (%)+0.25%.²¹

Criteria of metabolic syndrome

Two different sets of criteria were used separately for diagnosis of metabolic syndrome. Metabolic syndrome was defined, according to the criteria by NCEP ATP III,¹¹ as the presence of three or more risk factors or was defined, according to the criteria by IDF¹² with a slight modification, as the presence of two or more risk factors in addition to visceral obesity diagnosed as large WC. Risk factors included in the criteria are visceral obesity, high blood pressure, dyslipidemina (low HDL-C and/or high TGs), and hyperglycemia evaluated by HbA1c. The criterion for each risk factor was defined as follows: visceral obesity, WC ≥85 cm; high blood pressure, systolic blood pressure (SBP) ≥130 mmHg and/or DBP ≥85 mmHg; low HDL-C, HDL-C <40 mg/dL; high TGs, TGs ≥150 mg/dL; hyperglycemia, HbA1c ≥6.5%. Subjects receiving drug therapy for hypertension, dyslipidemia and diabetes were also included in the above definitions of high blood pressure, dyslipidemia, and hyperglycemia, respectively. Because the subjects of this study were patients with diabetes, they had at least one of the risk factors comprising metabolic syndrome.

Statistical analysis

Statistical analyses were performed using a computer software program (SPSS v. 16.0 J for Windows, Chicago IL). Categorical variables were compared between each pair of groups using the chi-squared test for independence. For continuous variables, means of each variable were compared among the groups by using analysis of variance (ANOVA) followed by the Scheffé F-test as a post hoc test in univariate analysis. In multivariate analysis, mean levels of each variable were compared by using analysis of covariance (ANCOVA) followed by the Student t-test after Bonferroni correction. Because TG levels did not show a normal distribution, they were used after log transformation in ANOVA and ANCOVA. In logistic regression analysis, crude and adjusted ORs for each risk factor or metabolic syndrome were calculated. Age, regular exercise, and drug therapy for diabetes were used as other explanatory variables or co-variables in multivariate analyses. In multivariate analyses of variables other than WC and metabolic syndrome, BMI was also added to the explanatory variables and co-variables. Probability (P) values less than 0.05 were defined as significant.

Results

Characteristics of the three smoker groups and overall subjects

Table 1 shows the characteristics of the nonsmoker, light smoker, and heavy smoker groups and overall subjects. Age tended to be younger with an increase in the amount of smoking. The proportions of subjects doing exercise regularly and subjects receiving drug therapy for diabetes, hypertension, or dyslipidemia tended to be lower with an increase in the amount of smoking. The HbA1c level tended to be higher with an increase in the amount of smoking. BMI was slightly but significantly higher in heavy smokers than in nonsmokers. The proportions of subjects showing large WC, high TGs, low HDL-C, dyslipidemia, and metabolic syndrome, diagnosed by using the criteria of NCEP ATP III or IDF, were significantly higher in heavy smokers than in nonsmokers. The proportion of subjects showing large WC was significantly lower in light smokers than in nonsmokers. The proportion of subjects showing low HDL-C tended to be higher with an increase in the amount of smoking.

Table 1.

Characteristics of the Groups of Nonsmokers, Light Smokers, and Heavy Smokers in Subjects with Diabetes

	Nonsmokers	Light smokers	Heavy smokers	Overall subjects
Number	1293	891	491	2675
Age (years)	55.4±7.3	53.3±7.7^**	51.8±7.6^**	54.1±7.6
Alcohol drinkers (%)	73.2	76.1	75.8	74.6
Regular exercise (%)	14.8	11.2^*	9.4^**	12.6
Therapy for diabetes (%)	61.9	54.8^**	49.1^**	57.2
Therapy for hypertension (%)	40.8	28.7^**	25.7^**	34.0
Therapy for dyslipidemia (%)	19.3	15.2^*	15.1^*	17.1
Hemoglobin A1c (%)	7.47±1.51	7.65±1.58^*	7.82±1.57^**	7.60±1.55
BMI (kg/m²)	25.55±3.89	25.23±4.05	26.12±4.28^*	25.55±4.03
Large WC (%)	63.7	58.9^*	72.1^**	63.7
High blood pressure (%)	79.2	75.0	76.4	77.3
High TGs (%)	44.8	47.5	56.0^**	47.7
Low HDL-C (%)	13.9	19.1^**	23.2^**	17.3
Dyslipidemia (%)	57.0	58.5	65.4^**	59.0
Metabolic syndrome by NCEP ATP III (%)	70.5	67.6	77.8^**	70.8
Metabolic syndrome by IDF (%)	59.8	55.1	68.0^**	59.7

Shown are number of subjects, percentage of subjects and mean with standard deviation of each variable. Symbols denote significant differences from nonsmokers (^* P<0.05; ^** P<0.01).

BMI, body mass index; WC, waist circumference; TGs, triglycerides; HDL-C, high-density lipoprotein cholesterol; NCEP ATP III, National Cholesterol Education Program Adult Treatment Panel III; IDF, International Diabetes Federation.

Comparison of the mean levels of variables comprising metabolic syndrome among the nonsmoker, light smoker, and heavy smoker groups in each drinker group

The mean levels of variables comprising metabolic syndrome were compared among the smoker groups in each drinker group (Table 2). In the nondrinker group, the log-transformed TGs level was significantly higher in heavy smokers than in nonsmokers, and HDL-C was significantly lower in light and heavy smokers than in nonsmokers. In the light drinker group, WC, SBP and DBP, log-transformed TGs, and HDL-C were not significantly different between nonsmokers and light smokers and between nonsmokers and heavy smokers. In the heavy drinker group, WC and DBP were significantly smaller and lower, respectively, in light smokers than in nonsmokers, whereas HDL-C was significantly lower in heavy smokers than in nonsmokers.

Table 2.

Comparison of Mean Levels of Each Component for Metabolic Syndrome Among Nonsmokers, Light Smokers, and Heavy Smokers in the Nondrinker, Light Drinker, and Heavy Drinker Groups

	Nondrinkers Smokers Non	Light	Heavy	Light drinkers Smokers Non	Light	Heavy	Heavy drinkers Smokers Non	Light	Heavy
WC (cm)
Univariate	88.6±0.6	89.2±0.8	93.8±1.1^**	87.8±0.5	87.3±0.7	89.1±1.2	88.6±0.4	87.4±0.4	89.8±0.6
Multivariate	89.6±0.6	88.5±0.7	92.2±1.0	87.9±0.5	87.0±0.7	88.8±1.2	89.0±0.4	87.2±0.4^**	89.1±0.5
SBP (mmHg)
Univariate	133.4±0.9	133.7±1.3	138.1±1.9	135.4±1.0	134.4±1.2	133.0±2.1	139.4±0.7	137.4±0.8	139.6±1.0
Multivariate	133.4±0.9	134.4±1.2	136.9±1.6	134.9±0.9	135.2±1.1	133.1±2.1	138.8±0.7	138.1±0.8	139.8±1.0
DBP (mmHg)
Univariate	79.1±0.6	80.1±0.9	81.7±1.4	80.9±0.7	78.8±0.8	77.6±1.4	83.4±0.5	81.3±0.5^**	83.2±0.7
Multivariate	79.5±0.6	80.3±0.8	80.1±1.0	81.0±0.6	78.7±0.8	76.9±1.4	83.6±0.4	81.6±0.5^*	82.4±0.6
Log[TGs (mg/dL)]
Univariate	2.12±0.01	2.18±0.02^*	2.26±0.03^**	2.12±0.02	2.15±0.02	2.05±0.03	2.17±0.01	2.17±0.01	2.23±0.02^*
Multivariate	2.13±0.01	2.18±0.02	2.23±0.02^**	2.12±0.02	2.14±0.02	2.03±0.03	2.18±0.01	2.18±0.01	2.21±0.02
HDL-C (mg/dL)
Univariate	49.6±0.7	45.4±0.7^**	42.7±1.0^**	52.6±0.8	49.3±1.1^*	50.3±1.6	56.2±0.6	55.0±0.6	52.1±0.8^**
Multivariate	49.6±0.6	45.1±0.8^**	43.2±1.0^**	52.6±0.8	49.2±1.0	50.9±1.7	56.0±0.5	54.7±0.6	53.0±0.8^*

Means with their standard errors of each variable are shown. In multivariate analysis, adjusted means were calculated using age, regular exercise, and drug therapy for diabetes as other explanatory co-variates. In addition, BMI was also adjusted for calculation of the means of variables other than waist circumference. Drug therapy for hypertension and dyslipidemia was further used as an explanatory covariate for calculation of adjusted means of blood pressure and blood lipids, respectively.

The P values were adjusted by three times each P value to test significance. Asterisks denote significant differences from nonsmokers in each drinking group (^* P<0.05; ^** P<0.01).

WC, waist circumference; SBP, systolic blood pressure; DBP, diastolic blood pressure; TGs, triglycerides; HDL-C, high-density lipoprotein cholesterol.

ORs of light or heavy smokers versus nonsmokers for risk factors comprising metabolic syndrome in each drinker group

Table 3 shows ORs of light or heavy smokers versus nonsmokers for each risk factor comprising metabolic syndrome. In the nondrinker group, ORs versus nonsmokers of heavy smokers for large WC, high TGs, low HDL-C, and dyslipidemia were significantly higher than the reference level of 1.00. In addition, light smokers showed significantly higher ORs for low HDL-C and dyslipidemia in the nondrinker group. In the light drinker group, ORs versus nonsmokers of light and heavy smokers for each component of metabolic syndrome were not significantly different from the reference level, except for a significantly higher OR for low HDL-C of light smokers and a significantly lower OR for high blood pressure of heavy smokers. In the heavy drinker group, ORs versus nonsmokers of light and heavy smokers for each component of metabolic syndrome were not significantly different from the reference level, except for a significantly lower OR for large WC of light smokers and significantly higher ORs for high TGs and low HDL-C of heavy smokers.

Table 3.

Odds Ratios of Smokers Versus Nonsmokers for Large Waist Circumference, High Blood Pressure, High Triglycerides, Low HDL-C, and Dyslipidemia in the Nondrinker, Light Drinker, and Heavy Drinker Groups

	Nondrinkers Smokers Non	Light	Heavy	Light drinkers Smokers Non	Light	Heavy	Heavy drinkers Smokers Non	Light	Heavy
Large WC
Univariate	1.00	0.98 (0.69–1.39)	2.03^** (1.27–3.24)	1.00	0.89 (0.61–1.30)	1.54 (0.81–2.91)	1.00	0.72^** (0.57–0.92)	1.27 (0.95–1.70)
Multivariate	1.00	0.79 (0.55–1.15)	1.64^* (1.01–2.68)	1.00	0.82 (0.56–1.22)	1.44 (0.75–2.75)	1.00	0.64^** (0.50–0.83)	1.03 (0.76–1.40)
High blood pressure
Univariate	1.00	0.92 (0.64–1.33)	1.20 (0.75–1.92)	1.00	0.67 (0.44–1.03)	0.57 (0.30–1.06)	1.00	0.72^* (0.53–0.97)	0.71^* (0.50–0.99)
Multivariate	1.00	1.00 (0.67–1.50)	1.02 (0.60–1.72)	1.00	0.73 (0.46–1.16)	0.46^* (0.23–0.92)	1.00	0.83 (0.61–1.14)	0.74 (0.52–1.07)
High TGs
Univariate	1.00	1.29 (0.92–1.82)	2.27^** (1.48–3.48)	1.00	1.21 (0.84–1.76)	0.61 (0.33–1.14)	1.00	1.00 (0.79–1.26)	1.54^** (1.18–2.02)
Multivariate	1.00	1.22 (0.85–1.76)	1.78^* (1.12–2.81)	1.00	1.14 (0.77–1.69)	0.53 (0.28–1.00)	1.00	0.98 (0.77–1.26)	1.35^* (1.01–1.80)
Low HDL-C
Univariate	1.00	1.83^** (1.24–2.70)	3.07^** (1.97–4.79)	1.00	1.87^* (1.16–3.02)	1.04 (0.46–2.36)	1.00	1.16 (0.80–1.70)	1.75^** (1.19–2.57)
Multivariate	1.00	1.97^** (1.32–2.95)	3.20^** (1.99–5.14)	1.00	1.86^* (1.13–3.04)	0.98 (0.42–2.26)	1.00	1.17 (0.80–1.70)	1.66^* (1.11–2.50)
Dyslipidemia
Univariate	1.00	1.37 (0.96–1.94)	2.24^** (1.41–3.55)	1.00	1.05 (0.72–1.52)	0.62 (0.35–1.11)	1.00	0.96 (0.75–1.21)	1.40^* (1.06–1.86)
Multivariate	1.00	1.47^* (1.01–2.14)	2.04^** (1.24–3.34)	1.00	1.01 (0.68–1.49)	0.57 (0.31–1.03)	1.00	0.99 (0.77–1.26)	1.32 (0.98–1.77)

Adjusted odds ratios with their 95% confidence intervals indicated in the parentheses are shown. In multivariate analysis, adjusted odds ratios were calculated using age, regular exercise, and drug therapy for diabetes as other explanatory variables. BMI was also used as an explanatory variable to calculate adjusted odds ratios for high blood pressure, high triglycerides, low HDL-C, and dyslipidemia.

Asterisks denote significantly lower or higher odds ratios compared with a reference level of 1.00 (^* P<0.05; ^** P<0.01).

WC, waist circumference; TGs, triglycerides; HDL-C, high-density lipoprotein cholesterol.

ORs of light or heavy smokers versus nonsmokers for metabolic syndrome in each drinker group

ORs of light or heavy smokers versus nonsmokers for metabolic syndrome diagnosed by the NCEP ATP III or IDF criteria are shown in Table 4. In the nondrinker group, heavy smokers showed significantly higher ORs versus nonsmokers for metabolic syndrome than the reference level of 1.00. On the other hand, in the light and heavy drinkers, the ORs versus nonsmokers of light and heavy smokers for metabolic syndrome were not significantly different from the reference level, except for significantly lower ORs of light smokers than the reference level in the heavy drinker group.

Table 4.

Odds Ratios of Smokers Versus Nonsmokers for Metabolic Syndrome in the Nondrinker, Light Drinker, and Heavy Drinker Groups

	Nondrinkers Smokers Non	Light	Heavy	Light drinkers Smokers Non	Light	Heavy	Heavy drinkers Smokers Non	Light	Heavy
Metabolic syndrome (NCEP ATP III)
Univariate	1.00	1.16 (0.81–1.66)	2.93^** (1.73–4.96)	1.00	0.81 (0.55–1.21)	0.88 (0.48–1.63)	1.00	0.75^* (0.58–0.98)	1.19 (0.86–1.63)
Multivariate	1.00	1.03 (0.71–1.49)	2.47^** (1.43–4.25)	1.00	0.76 (0.51–1.14)	0.82 (0.44–1.54)	1.00	0.69^** (0.53–0.90)	1.00 (0.72–1.39)
Metabolic syndrome (IDF)
Univariate	1.00	1.02 (0.72–1.44)	2.27^** (1.44–3.59)	1.00	0.90 (0.62–1.31)	1.22 (0.68–2.22)	1.00	0.71^** (0.56–0.90)	1.20 (0.90–1.59)
Multivariate	1.00	1.09 (0.66–1.78)	1.81^* (1.13–2.92)	1.00	0.83 (0.57–1.22)	1.15 (0.63–2.10)	1.00	0.64^** (0.50–0.82)	0.96 (0.72–1.30)

Asterisks denote significantly lower or higher odds ratios compared with a reference level of 1.00 (^* P<0.05; ^** P<0.01).

NCEP ATP III, National Cholesterol Education Program Adult Treatment Panel III; IDF, International Diabetes Federation.

ORs versus nondrinkers who were nonsmokers for metabolic syndrome in each smoking group of light and heavy drinkers

As shown in Table 5, the ORs versus nondrinkers who were nonsmokers for metabolic syndrome diagnosed by the NCEP ATP III or IDF criteria were significantly higher than a reference level of 1.00 in the heavy drinkers who were nonsmokers and in those who were heavy smokers but were not significantly different from the reference level in heavy drinkers who were light smokers. The ORs versus nondrinkers who were nonsmokers for metabolic syndrome diagnosed by the NCEP ATP III criteria were significantly higher than the reference level in light drinkers who were nonsmokers (Table 5).

Table 5.

Odds Ratios Versus Nondrinkers Who Were Nonsmokers for Metabolic Syndrome in Each Smoking Group of Light and Heavy Drinkers

	Nondrinkers Smokers Non	Light drinkers Smokers Non	Light	Heavy	Heavy drinkers Smokers Non	Light	Heavy
Metabolic syndrome (NCEP ATP III)
Univariate	1.00	1.42^* (1.02–1.98)	1.16 (0.80–1.68)	1.25 (0.69–2.28)	1.72^** (1.30–2.28)	1.30 (0.97–1.74)	2.04^** (1.45–2.88)
Multivariate	1.00	1.50^* (1.07–2.11)	1.11 (0.76–1.62)	1.19 (0.65–2.19)	1.77^** (1.33–2.35)	1.27 (0.94–1.70)	1.84^** (1.29–2.61)
Metabolic syndrome (IDF)
Univariate	1.00	1.23 (0.90–1.69)	1.11 (0.77–1.59)	1.51 (0.84–2.70)	1.41^* (1.08–1.83)	1.00 (0.76–1.32)	1.68^** (1.23–2.30)
Multivariate	1.00	1.30 (0.94–1.80)	1.06 (0.73–1.52)	1.43 (0.79–2.59)	1.46^** (1.11–1.91)	0.96 (0.72–1.27)	1.47^* (1.06–2.03)

Asterisks denote significantly lower or higher odds ratios compared with a reference level of 1.00 (^* P<0.05; ^** P<0.01).

NCEP ATP III, National Cholesterol Education Program's Adult Treatment Panel III; IDF, International Diabetes Federation.

Discussion

This study, for the first time, demonstrated that there are diverse associations between smoking and metabolic syndrome in patients with diabetes, depending on the amount of smoking and history of alcohol drinking. In patients with diabetes, heavy smoking (>20 cigarettes per day) was positively associated with metabolic syndrome in nondrinking subjects but not in drinking subjects, whereas light smoking (≤20 cigarettes per day) was inversely associated with metabolic syndrome in the heavy drinker group. The results for the relationships between smoking and metabolic syndrome were similar when the different criteria proposed by NCEP ATP III and IDF were used for diagnosis of metabolic syndrome. Thus, the relationship between smoking and metabolic syndrome is modified by alcohol drinking, which is likely to counteract the harmful effects of smoking on the risk factors comprising metabolic syndrome, especially the lowering effect on HDL-C. In fact, the ORs of light and heavy smokers versus nonsmokers for low HDL-C in the heavy drinker group were lower than the corresponding ORs in the nondrinker group (Table 3).

Previous studies demonstrated that active smoking increases the risk of metabolic syndrome in general populations.^1

–4 The present finding of a positive association of smoking with metabolic syndrome in nondrinkers suggests that the relationship between smoking and metabolic syndrome is similar in persons with diabetes and those without diabetes. Because both smoking and diabetes are potent risk factors for CVD, it is obvious that no smoking is strongly recommended for patients with diabetes. Increased risk of metabolic syndrome by smoking is thought to be, in part, involved in acceleration of atherosclerotic progression in smokers with diabetes. Moreover, HbA1c levels tended to be higher with an increase in the amount of smoking (Table 1), and this agrees with the results of a systematic review with meta-analysis of 25 prospective studies showing that the risk of type 2 diabetes was higher in smokers than in nonsmokers.²²

It is known that smokers generally have lower HDL-C and higher TG levels than nonsmokers,⁶ which are explained by deterioration of lipid metabolism via decreased lipoprotein lipase activity.²³ Similar relationships of smoking with lipid levels were found in patients with diabetes. In addition, in the nondrinker group, WC tended to be larger in heavy smokers than in nonsmokers. Therefore, proneness to dyslipidemia, such as low HDL cholesterolemia and hypertriglyceridemia, and visceral obesity, reflected by large WC, are thought to contribute to the higher risk of metabolic syndrome in heavy smokers with diabetes.

It is known that body weight of smokers tends to be less than that of nonsmokers, and this may be explained by the actions of nicotine, such as an increase in energy expenditure and reduction of appetite. However, as shown in Table 1, BMI and WC were significantly higher and larger, respectively, in heavy smokers with diabetes than in nonsmokers with diabetes. This dissociation may result from risky behaviors of smokers with diabetes, because patients with diabetes are strongly recommended not to smoke to prevent complications of CVD. In fact, the proportions of subjects doing exercise regularly and receiving therapy for diabetes were significantly lower in heavy smokers than in nonsmokers (Table 1). Physical activity and diet are strongly related to visceral obesity, a central risk factor for metabolic syndrome. Therefore, the association between heavy smoking and metabolic syndrome may be partly explained by a confounding with low level of physical activity and unhealthy diet frequently encountered among heavy smokers.

In the heavy drinker group, light smokers showed an inverse association with metabolic syndrome, which was contributed to by smaller WC and lower DBP in light smokers than in nonsmokers. A possible reason for the above inverse association observed in the heavy drinker group, but not in the nondrinker and light drinker groups, is that the lowering effect of smoking on HDL-C, a component of metabolic syndrome, was canceled by the HDL-C–elevating effect of alcohol. In fact, HDL-C (mean±standard error after adjustment for age, BMI, smoking, regular exercise, and drug therapy for diabetes and dyslipidemia) was significantly higher (P<0.01) in the heavy drinker group (54.9±0.3 mg/dL) than in the nondrinker group (47.7±0.5 mg/dL) and light drinker group (50.6±0.6 mg/dL).

As shown in Tables 4 and 5, the OR versus nondrinkers who were nonsmokers for metabolic syndrome was highest in nondrinkers who were heavy smokers and was second highest in heavy drinkers who were heavy smokers. Therefore, both heavy smoking and heavy drinking are thought to contribute to increase in the risk of metabolic syndrome. The significantly high ORs for metabolic syndrome found in heavy drinkers who were nonsmokers and in those who were heavy smokers could be explained by the positive associations of heavy drinking with high blood pressure and high TGs (Table 2). Smaller WC in heavy drinkers who were light smokers than that in nondrinkers who were nonsmokers as shown in Table 2 may be responsible for no significant difference from the reference level of the OR versus nondrinkers who were nonsmokers for metabolic syndrome in heavy drinkers who were light smokers.

There are limitations of this study. First, diabetes was diagnosed by HbA1c level and history of drug therapy for diabetes in the questionnaire. Thus, both patients with type 1 and type 2 diabetes were included in the subjects. However, the type of diabetes for most of the subjects is expected to be type 2 because the prevalence of type 2 diabetes is speculated to be >100 times higher than that of type 1 diabetes in middle-aged Japanese men according to statistics.²⁴ Unfortunately, the reliability and validity of the questionnaires were unknown, and there is a possibility of information bias in this study. Moreover, information on duration of diabetes, which possibly confounds the relation of smoking to metabolic syndrome in patients with diabetes, was not included in the database used in this study. Smokers were divided into only two groups of light and heavy smokers, because the population size of smokers was not large enough for more detailed classification of smokers. In addition, no information on ex-smokers was available, and ex-smokers were included in the nonsmoker group. The risk of metabolic syndrome for ex-smokers (smoking cessation within 1 year) has been reported to be comparable to the risk for nonsmokers.²⁵ The subjects were Japanese men with diabetes, and further studies using data for subjects of different race and/or ethnicity and female subjects are needed to confirm the findings of the present study. Finally, this study is cross-sectional in its design, and further studies are also needed to clarify the causal relationship between smoking and metabolic syndrome in patients with diabetes.

In conclusion, in nondrinkers with diabetes, there is a positive association between active smoking (>20 cigarettes per day) and metabolic syndrome with an OR for metabolic syndrome being about 2, which is mainly due to higher risks of dyslipidemia and visceral obesity in smokers than in nonsmokers. Alcohol drinking is likely to modify the relationship between smoking and metabolic syndrome.

Footnotes

Acknowledgments

This work was supported by a Grant-in-Aid for Scientific Research from the Japan Society for the Promotion of Science (No. 24390171).

Author Disclosure Statement

No potential conflicts of interest relevant to this article were reported.

References

Nakanishi

, Takatorige

, Suzuki

. Cigarette smoking and the risk of the metabolic syndrome in middle-aged Japanese male office workers. Ind Health, 2005; 43:295–301.

Ishizaka

, Ishizaka

, Toda

, et al. Association between cigarette smoking, metabolic syndrome, and carotid arteriosclerosis in Japanese individuals. Atherosclerosis, 2005; 181:381–388.

Wannamethee

, Shaper

, Whincup

. Modifiable lifestyle factors and the metabolic syndrome in older men: Effects of lifestyle changes. J Am Geriatr Soc, 2006; 54:1909–1914.

Onat

, Özhan

, Esen

, et al. Prospective epidemiologic evidence of a “protective” effect of smoking on metabolic syndrome and diabetes among Turkish women—without associated overall health benefit. Atherosclerosis, 2007; 193:380–388.

Sun

, Liu

, Ning

. Active smoking and risk of metabolic syndrome: A meta-analysis of prospective studies. PLoS One, 2012; 7:e47791.

Craig

, Palomaki

, Haddow

. Cigarette smoking and serum lipid and lipoprotein concentrations: An analysis of published data. Br Med J, 1989; 298:784–788.

Isomaa

, Almgren

, Tuomi

, et al. Cardiovascular morbidity and mortality associated with the metabolic syndrome. Diabetes Care, 2001; 24:683–689.

Hillier

, Rizzo

, Pedula

, et al. Increased mortality associated with the metabolic syndrome in older women with diabetes. Diabetes Care, 2005; 28:2258–2260.

Saely

, Aczel

, Marte

, et al. The metabolic syndrome, insulin resistance, and cardiovascular risk in diabetic and nondiabetic patients. J Clin Endocrinol Metab, 2005; 90:5698–5703.

10.

Alexander

, Landsman

, Teutsch

, et al. NCEP-defined metabolic syndrome, diabetes, and prevalence of coronary heart disease among NHANES III participants age 50 years and older. Diabetes, 2003; 52:1210–1214.

11.

Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults. Executive Summary of the Third Report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III). JAMA, 2001; 285:2486–2497.

12.

Alberti

, Zimmet

, Shaw

; IDF Epidemiology Task Force Consensus Group. The metabolic syndrome—a new worldwide definition. Lancet, 2005; 366:1059–1062.

13.

Corrao

, Rubbiati

, Bagnardi

, et al. Alcohol and coronary heart disease: A meta-analysis. Addiction, 2000; 95:1505–1523.

14.

Ellison

, Zhang

, Qureshi

, et al.; Investigators of the NHLBI Family Heart Study. Lifestyle determinants of high-density lipoprotein cholesterol: The National Heart, Lung, and Blood Institute Family Heart Study. Am Heart J, 2004; 147:529–535.

15.

Alkerwi

, Boutsen

, Vaillant

, et al. Alcohol consumption and the prevalence of metabolic syndrome: A meta-analysis of observational studies. Atherosclerosis, 2009; 204:624–635.

16.

Wakabayashi

. Association between alcohol drinking and metabolic syndrome in Japanese male workers with diabetes mellitus. J Atheroscler Thromb, 2011; 18:684–692.

17.

Anonymous; American Diabetes Association. Diagnosis and classification of diabetes mellitus. Diabetes Care, 2010; 33(Suppl 1):S62–S69.

18.

Ogihara

, Kikuchi

, Matsuoka

, et al.; Japanese Society of Hypertension Committee. The Japanese Society of Hypertension Guidelines for the Management of Hypertension (JSH 2009). Hypertens Res, 2009; 32:3–107.

19.

Mancia

, Fagard

, Narkiewicz

, et al. 2013 ESH/ESC Guidelines for the management of arterial hypertension: The Task Force for the management of arterial hypertension of the European Society of Hypertension (ESH) and of the European Society of Cardiology (ESC). J Hypertens, 2013; 31:1281–1357.

20.

Anonymous. Metabolic syndrome—Definition and diagnostic criteria in Japan. J Jpn Soc Int Med, 2005; 94:794–809 (in Japanese).

21.

Kashiwagi

, Kasuga

, Araki

, et al.; Committee on the Standardization of Diabetes Mellitus-Related Laboratory Testing of Japan Diabetes Society (JDS). International clinical harmonization of glycated hemoglobin in Japan: From Japan Diabetes Society to National Glycohemoglobin Standardization Program values. Diabetol Int, 2012; 3:8–10.

22.

Willi

, Bodenmann

, Ghali

, et al. Active smoking and the risk of type 2 diabetes: A systematic review and meta-analysis. JAMA, 2007; 298:2654–2664.

23.

Freeman

, Caslake

, Griffin

, et al. The effect of smoking on post-heparin lipoprotein and hepatic lipase, cholesteryl ester transfer protein and lecithin:cholesterol acyl transferase activities in human plasma. Eur J Clin Invest, 1998; 28:584–591.

24.

Anonymous. International Diabetes Federation Diabetes Atlas, 4th ed. 2009. Available at: www.diabetesatlas.org

25.

Kawada

, Otsuka

, Inagaki

, et al. Association of smoking status, insulin resistance, body mass index, and metabolic syndrome in workers: A 1-year follow-up study. Obes Res Clin Pract, 2010; 4; e163–e169.