Smoking is a leading risk factor for noncommunicable diseases (NCDs), such as cardiovascular disease [
The relationship between smoking and type 2 diabetes mellitus (T2DM) is unclear. Although most epidemiological evidence has demonstrated a positive association between smoking and T2DM [
The Risk Evaluation of Cancers in Chinese Diabetic Individuals: a community-based longitudinal (REACTION) study was a multicenter prospective observational study of adults aged 40 years or older. The primary purpose of the study was to assess the relationship between abnormal glucose metabolism and the risk of cancer in the Chinese population. Details of the study design have been published previously [
All participants were interviewed by qualified health employees using face-to-face standardized questionnaires. Data collected included educational level, smoking status (including duration of smoking history or since quitting smoking), alcohol consumption, physical activity, and medical history (including current medications and diseases such as diabetes mellitus, hypertension, and stroke). A total of 8196 adults aged ≥40 years old were included in this analysis. Enrolled subjects who did not complete oral glucose tolerance testing (OGTT) or provided incomplete questionnaire information were excluded from the analysis. According to the theory of obesity paradox, participants with a
All included individuals were categorized into nonsmokers, exsmokers, and current smokers. Those who had never smoked or smoked less than 100 cigarettes in their lifetime were considered nonsmokers, those who had regularly smoked previously but had quit smoking >1 year prior to the interview were considered exsmokers, and those who currently smoked at least one cigarette per day or seven cigarettes per week on a regular basis, as well as those who had been quitting smoking <1 year prior to the interview were defined as current smokers. Exsmokers were asked how long ago they had stopped smoking. Current smokers were asked the age at which they started smoking, years of smoking, and number of cigarettes smoked per day. Smoking amount was defined according to the number of cigarettes per day (light: 1–19; heavy: ≥20) or pack-years (light: 1–19; heavy: ≥20). Pack-years were calculated as follows: packs (
Educational level was classified as low (no education, primary education, or junior middle education), medium (high-school education), or high (college education). Regular leisure-time physical activity was defined as participation in moderate or vigorous activity for more than 30 minutes per day at least three days a week. The diagnosis of nonfatal stroke was based on subjects’ self-reporting and previous medical history. All subjects who stated a history of stroke were required to state the hospital where they were diagnosed or provide evidence of disease diagnosis.
Anthropometric measurements including height, weight, waist circumference (WC), and hip circumference were performed using a standard protocol. BMI was calculated as weight divided by height squared. The waist to hip ratio (WHR) was calculated as the ratio of waist and hip circumference.
All participants underwent an OGTT and provided blood samples which were tested for fasting plasma glucose (FPG), and 2-hour postprandial plasma glucose (2hPG), fasting insulin (FINS), hemoglobin A1c (HbA1c), total cholesterol (TC), triglyceride (TG), high-density lipoprotein cholesterol (HDL-C), and low-density lipoprotein cholesterol (LDL-C) according to China National Laboratory Accreditation [
All participants were required to fast for at least 10 hours before undergoing a 75 g OGTT. Diagnosis of normal glucose tolerance (NGT), prediabetes, and T2DM was based on the 2017 T2DM diagnostic criteria of the Chinese Diabetes Society (CDS) [
All statistical analyses were conducted using SPSS version 23.0 (SPSS Inc., Chicago, IL). We first conducted analysis of variance (ANOVA), the Kruskal-Wallis test (as appropriate), or the
The clinical baseline characteristics of the total study population and the subgroups studied are summarized in Table
Characteristics of the population according to the status of glucose metabolism (
Variables | NGT (A) |
Prediabetes (B) |
T2DM (C) |
|||
---|---|---|---|---|---|---|
Age (years) | ||||||
Sex, male/female ( |
1396/3003 | 637/1227 | 741/1192 | 0.177 | ||
BMI (kg/m2) | ||||||
WC (cm) | ||||||
WHR | ||||||
FPG (mmol/L) | ||||||
2hPG (mmol/L) | ||||||
HbA1c (%) | ||||||
TG (mmol/L) | 1.09 (0.82-1.54) | 1.32 (0.96-1.85) | 1.36 (0.98-2.00) | |||
TC (mmol/L) | ||||||
LDL-C (mmol/L) | 0.082 | |||||
HDL-C (mmol/L) | ||||||
FINS ( |
5.50 (4.10-7.50) | 6.40 (4.60-8.70) | 6.90 (4.70-10.10) | |||
HOMA-IR | 1.26 (0.92-1.75) | 1.60 (1.12-2.22) | 2.15 (1.39-3.35) | |||
HOMA- |
57.02 (49.41-92.06) | 51.26 (43.63-86.06) | 43.52 (27.02-69.69) | |||
SBP (mmHg) | ||||||
DBP (mmHg) | 0.080 | |||||
Educational level | ||||||
Low (%) | 3216 (73.1%) | 1328 (71.2%) | 1291 (66.8%) | — | — | — |
Medium (%) | 960 (21.8%) | 389 (20.9%) | 439 (22.7%) | — | — | — |
High (%) | 223 (5.1%) | 147 (7.9%) | 203 (10.5%) | — | — | — |
Current drinker (%) | 1101 (25.0%) | 487 (26.1%) | 469 (24.3%) | 0.411 | — | — |
Physical activity (%) | 636 (14.5%) | 315 (16.8%) | 388 (20.1%) | |||
Family history of DM (%) | 477 (10.8%) | 236 (12.7%) | 352 (18.2%) | 0.114 | ||
Smoking status | ||||||
Never-smoker (%) | 3443 (78.3%) | 1495 (80.2%) | 1524 (78.8%) | — | — | — |
Exsmoker (%) | 124 (2.8%) | 74 (4.0%) | 94 (4.9%) | — | — | — |
Current smoker (%) | 832 (18.9%) | 295 (15.8%) | 315 (16.3%) | — | — | — |
Nonfatal stroke (%) | 65 (1.5%) | 41 (2.2%) | 82 (4.2%) | 0.129 |
Data are
After adjustment for age, BMI, gender, educational level, physical activity, alcohol consumption, family history of diabetes, SBP, DBP, TG, TC, HDL-C, and LDL-C, current smokers were associated with a 23% (OR: 0.77, 95% CI: 0.63–0.93) and 20% (OR: 0.80, 95% CI: 0.66–0.98) decreased risk of prediabetes and T2DM compared with nonsmokers, respectively.
In addition, light smokers (1–19 cigarettes/day or 1–19 pack-years) exhibited a lower risk of having T2DM, while no significant association with risk of T2DM was found for heavy smokers and nonsmokers (Table
Association of smoking status with T2DM (
Model 1 | Model 2 | Model 3 | ||||
---|---|---|---|---|---|---|
OR | OR | OR | ||||
Smoking status | ||||||
Never-smoker | 1 | 1 | 1 | |||
Exsmoker | 0.80 (0.71-1.33) | 0.848 | 0.97 (0.70-1.33) | 0.847 | ||
Current smoker | ||||||
Daily consumption | ||||||
Never-smoker | 1 | 1 | 1 | |||
1-19 cigarettes/day | 0.92 (0.76-1.12) | 0.419 | ||||
≥20 cigarettes/day | 0.80 (0.63-1.01) | 0.059 | 0.79 (0.62-1.01) | 0.057 | ||
Smoking duration | ||||||
Never-smoker | 1 | 1 | 1 | |||
<20 years | 0.87 (0.70-1.09) | 0.222 | ||||
≥20 years | 1.01 (0.87-1.18) | 0.870 | 0.87 (0.71-1.06) | 0.170 | 0.84 (0.68-1.04) | 0.106 |
Pack-year | ||||||
Never-smoker | 1 | 1 | 1 | |||
1-19 pack-years | ||||||
≥20 pack-years | 0.91 (0.75-1.10) | 0.307 | 0.82 (0.65-1.04) | 0.103 | 0.82 (0.65-1.05) | 0.112 |
Model 1: no adjusted variables. Model 2: adjusted for age, BMI, gender, educational level, physical activity, alcohol consumption, and family history of diabetes. Model 3: adjusted for Model 2 plus SBP, DBP, TG, TC, HDL-C, and LDL-C. The sample size was 6332 in the analysis of the subgroup of smoking status.
In men, stratification by normal weight (
Association between cigarette smoking and T2DM in men stratified by obesity (
Subgroup | Model 1 | Model 2 | Model 3 | |||
---|---|---|---|---|---|---|
OR | OR | OR | ||||
Never-smoker | 1 | 1 | 1 | |||
1-19 pack-years | ||||||
≥20 pack-years | ||||||
Never-smoker | 1 | 1 | 1 | |||
1-19 pack-years | 0.88 (0.60-1.29) | 0.524 | 0.87 (0.59-1.27) | 0.467 | ||
≥20 pack-years | 0.74 (0.54-1.03) | 0.072 | 1.03 (0.72-1.48) | 0.856 | 1.01 (0.71-1.44) | 0.959 |
Never-smoker | 1 | 1 | 1 | |||
1-19 pack-years | ||||||
≥20 pack-years | ||||||
Never-smoker | 1 | 1 | 1 | |||
1-19 pack-years | 0.80 (0.56-1.14) | 0.223 | 0.81 (0.57-1.16) | 0.257 | ||
≥20 pack-years | 0.73 (0.52-1.02) | 0.065 | 0.72 (0.51-1.01) | 0.060 | ||
Never-smoker | 1 | 1 | 1 | |||
1-19 pack-years | ||||||
≥20 pack-years | ||||||
Never-smoker | 1 | 1 | 1 | |||
1-19 pack-years | 0.72 (0.38-1.37) | 0.312 | 0.85 (0.43-1.69) | 0.639 | 0.79 (0.37-1.68) | 0.534 |
≥20 pack-years | 0.47 (0.22-1.03) | 0.059 | 0.45 (0.20-1.01) | 0.052 |
Model 1: no adjusted variables. Model 2: adjusted for age, BMI, gender, educational level, physical activity, alcohol consumption, and family history of diabetes. Model 3: adjusted for Model 2 plus SBP, DBP, TG, TC, HDL-C, and LDL-C.
Overall adiposity was defined as
Association between smoking status and overweight/obesity (
Individuals with HOMA-IR in the highest quartile were defined as insulin resistant. Using nonsmokers as the reference group, multivariable regression analysis in men showed that heavy smokers with
Association between smoking status and insulin resistance in men stratified by BMI (
The prevalence of self-report stroke was 2.3% in our study. No significant difference was observed in the prevalence of stroke between men and women (2.4% vs. 2.2%,
Interaction between diabetes and smoking in relation to stroke (
Smoking | T2DM | Model 1 | Model 2 | Model 3 | |||
---|---|---|---|---|---|---|---|
No | No | 1 | 1 | 1 | |||
No | Yes | 1.81 (1.28-2.55) | 0.001 | 1.91 (1.35-2.70) | <0.001 | 1.88 (1.33-2.66) | <0.001 |
Yes | No | 1.90 (1.06-3.43) | 0.032 | 1.87 (1.02-3.45) | 0.042 | 1.85 (1.01-3.41) | 0.047 |
Yes | Yes | 2.50 (1.20-5.21) | 0.014 | 2.67 (1.27-5.65) | 0.010 | 2.64 (1.25-5.57) | 0.011 |
Model 1: no adjusted variables. Model 2: adjusted for age, BMI, gender, educational level, physical activity, and alcohol consumption. Model 3: adjusted for Model 2 plus SBP, DBP, TG, TC, HDL-C, and LDL-C.
Smoking is an established major risk factor for various chronic noncommunicable diseases [
Similar to the studies described above, our findings show that cigarette smoking was associated with decreased odds of T2DM in men with
Given the different relationships identified between smoking and T2DM in different BMI/WC groups among men, we explored the relationship between smoking and obesity in the present. Logistic regression revealed significant negative associations between smoking and obesity in males, but no such association was found in women. Similarly, a previous study in Japan found that male current smokers had a lower OR for obesity compared with nonsmokers [
The inconsistent results between smoking and T2DM/obesity identified in our study were also found between men and women, as our data show that there was no evidence for an association between smoking and T2DM/obesity in women. The results of the Japanese study described above also showed that there were no differences in obesity between current smokers and nonsmokers [
Differences between our findings and those reported in several previous studies might in part be attributable to the following factors. First, the BMI cutoff points to define overweight and obesity were different in different studies. In addition, some studies used measures other than BMI to define obesity, including WC, WHR, visceral fat thickness, fat percentage, and fat mass. These differences in the method used to estimate obesity may have led to differences in the results. Second, the participants enrolled in the present study were all Chinese, so we used the CDS diabetes diagnostic criteria. The FPG levels for IFG were higher in the 2017 CDS diagnostic criteria compared with the diagnostic criteria of the American Diabetes Association used in other studies. Third, some other studies used only self-reported data, while our study combined medical histories and standard OGTT results to diagnose T2DM. Finally, the inclusion of different patient populations in the various studies might have contributed to population bias, thus affecting the study conclusions to some extent. Further studies of the relationship between smoking and obesity in patients with diabetes and the possible mechanism responsible for such a relationship are therefore required.
Diabetes is an established major risk factor for stroke [
This study had several limitations. First, although we enrolled 5422 women, there were only 87 current smokers in this group. The limited data might have affected the reliability of our conclusions for women. Further studies with a focus on women are thus required. Second, 27.1% of T2DM patients had received antidiabetic drugs, which might have affected the accuracy of the measurement of insulin levels and further affect HOMA-IR. Third, the diagnostic criteria for stroke were assessed based on patient self-reported data and medical history (although they provided evidence of disease diagnosis), which might have resulted in potential bias in the prevalence of stroke. Fourth, all subjects were of Chinese ethnicity, meaning that the results of the study might not be applicable to other populations. Finally, the use of a cross-sectional design means that it is not possible to infer a causal relationship between smoking status and prediabetes/T2DM. Prospective studies with larger sample sizes in specific populations are thus needed to clarify the relationship between smoking behavior and diabetes, related complications, and total morality.
The results of this cross-sectional study indicate that smoking is negatively associated with IGR/T2DM in Chinese men with normal body weight, although no significant association was found for overweight/obese Chinese men or for women. In addition, smoking and T2DM had a combined positive correlation with stroke after adjusting for confounders (including BMI). Therefore, smoking cessation should be suggested to all current smokers, regardless of BMI.
The data are held in a secure, confidential database, which can only be assessed by members of the REACTION group. The REACTION has a website (
No potential conflicts of interest relevant to this article were reported.
Su Wang and Jie Chen contributed equally to this work.
The REACTION study was supported by the grants from the Chinese Society of Endocrinology, and this work was supported in part by the National Natural Science Foundation of China (81870542) and the Special Foundation for “Top Six Talents” High Level Talent Selection and Training Foundation of Jiangsu Province (WNS-044). We thank Clare Cox, PhD, from Liwen Bianji, Edanz Editing China (
Supplementary Table 1: characteristics of the population according to the status of glucose metabolism (