High‐Intensity Cigarette Smoking Is Associated With Incident Diabetes Mellitus In Black Adults: The Jackson Heart Study
Background Previous reports on whether smoking is associated with insulin resistance and diabetes mellitus have yielded inconsistent findings. We aimed to evaluate the relationship between cigarette smoking and incident diabetes mellitus in the Jackson Heart Study.
Methods and Results Jackson Heart Study participants enrolled at baseline without prevalent diabetes mellitus (n=2991) were classified by self‐report as current smokers, past smokers (smoked ≥400 cigarettes/life and no longer smoking), or never smokers. We quantified smoking intensity by number of cigarettes smoked daily; we considered ≥20 cigarettes per day (1 pack) “high‐intensity.” We defined diabetes mellitus as fasting glucose ≥126 mg/dL, hemoglobin A1c ≥6.5% or International Federation of Clinical Chemistry units HbA1c 48 mmol/mol, or use of diabetes mellitus medication. We estimated the adjusted associations of smoking status, intensity, and dose (pack‐years) with incident diabetes mellitus using Poisson regression models. At baseline there were 361 baseline current (1–10 cigarettes per day [n=242]; ≥20 [n=119]), 502 past, and 2128 never smokers. From Visit 1 to Visit 3 (mean 8.0±0.9 years), 479 participants developed incident diabetes mellitus. After adjustment for covariates, baseline current smokers who smoked less than a pack/d and past smokers had similar rates of incident diabetes mellitus compared with never smokers (incidence rate ratios 1.04, 95% confidence interval, 0.69–1.58 and 1.08, 95% confidence interval, 0.82–1.42, respectively). Baseline current high‐intensity smokers had a 79% (95% confidence interval, 1.14–2.81) higher incidence of diabetes mellitus compared with never smokers. Smoking dose (per 10 pack‐years) was also associated with a higher incidence of diabetes mellitus (incidence rate ratios 1.10, 95% confidence interval, 1.03–1.19) in adjusted models.
Conclusions High‐intensity cigarette smoking and smoking pack‐years are associated with an increased risk of developing diabetes mellitus in blacks.
What Is New?
Current high‐intensity smoking (≥1 pack cigarettes per day) and smoking dose (pack‐years) are associated with incident diabetes mellitus in blacks.
Smoking intensity and dose were associated with impaired pancreatic beta cell function (homeostatic model assessment of pancreatic beta cell function).
What Are the Clinical Implications?
Although smoking cessation should be encouraged for everyone, certain high‐risk groups such as blacks who are disproportionately affected by diabetes mellitus should be targeted for cessation strategies.
Cigarette smoking is a major cause of preventable disease and mortality worldwide, and the impacts of smoking on cancer and cardiovascular disease are well established.1, 2, 3 Smoking also may be an independent risk factor for insulin resistance and diabetes mellitus; however, there have been discrepant findings from previous studies examining these relationships.4, 5, 6 Furthermore, despite evidence that smoking is associated with incident diabetes mellitus in certain ethnic groups, little is known about the association of cigarette smoking with diabetes mellitus in blacks, who are disproportionately affected by obesity and diabetes mellitus.7, 8
Racial/ethnic disparities in smoking behaviors have been described, including longer smoking durations for blacks compared with whites and lower cessation rates in blacks compared with both whites and Hispanics.9 In the NHANES II (National Health and Nutrition Examination Study), smoking rates were significantly higher in newly diagnosed black versus white individuals with diabetes mellitus (42% versus 29%, respectively).10 Behavioral smoking differences, coupled with a higher risk of other known diabetes mellitus risk factors, may augment the risk of diabetes mellitus in blacks who smoke.
Despite evidence suggesting a causal role for smoking predisposing to diabetes mellitus, the potential mechanisms that mediate this relationship are not well known. It has been hypothesized that smoking‐induced inflammation or obesity may contribute to diabetes mellitus onset.11, 12, 13 Although smoking has been associated with reduced body mass index (BMI), smoking is positively associated with visceral adiposity, a known risk factor for diabetes mellitus.14 We aimed to examine the associations of cigarette smoking status, intensity, and dose with incident diabetes mellitus in a large, prospective black cohort, the Jackson Heart Study (JHS). We also evaluated the influence of measures of adiposity (body mass index [BMI] and waist circumference [WC]), and inflammation (high‐sensitivity C‐reactive protein [hs‐CRP]) in this relationship.
Research Design and Methods
The JHS is a large, prospective cohort study of 5301 blacks aged 21 to 84 years who were recruited from the tri‐county area surrounding Jackson, MS from 2000 to 2004. We collected data at 3 participant examinations; we completed Visit 3 from 2009 to 2013. The 3 participating institutions (University of Mississippi Medical Center, Tougaloo College, and Jackson State University) approved the study and the University of Mississippi Medical Center Institutional Review Board approved the study. Each participant provided written, informed consent. The data, analytic methods, and study materials can be made available to other researchers for purposes of reproducing the results or replicating the procedure by following the JHS publications procedures and data use agreements.15
We included participants with available follow‐up data including diabetes mellitus outcome measures and self‐reported smoking status at Visit 3. We excluded participants with prevalent diabetes mellitus at baseline (Visit 1, n=733). We also excluded participants with relevant missing variables at Visit 3 (n=435).
Assessment of Smoking Status, Intensity, and Duration
We obtained self‐reported cigarette smoking status via questionnaire at baseline. We defined participants who smoked >400 cigarettes in their lifetime as “ever smokers.” We classified those who did not as “never smokers.” We classified participants who gave a positive response to the question “Do you now smoke cigarettes” as “baseline current smokers.” Participants who were classified as “ever” smokers who did not smoke for at least 30 days at the time of the examination were classified as “past” smokers. Further information related to number of cigarettes smoked daily also was collected. We defined smoking intensity as number of cigarettes smoked per day. We defined high‐intensity smoking as smoking more than 20 cigarettes daily based on previous published literature.16 We measured cumulative dose of smoking in “pack‐years,” which we calculated by multiplying the number of packs of cigarettes per day by the number of years smoking. One participant who identified as a current smoker did not report the number of cigarettes smoked daily and therefore was not included in analyses of smoking intensity or dose relationships.
Diabetes Mellitus Ascertainment
We defined diabetes mellitus as fasting plasma glucose ≥126 mg/dL, hemoglobin A1c (HbA1c) ≥6.5% or International Federation of Clinical Chemistry units HbA1c 48 mmol/mol, or use of diabetes mellitus medications within 2 weeks before the examination. We defined incident diabetes mellitus as presence of diabetes mellitus at Visit 3 in those participants who did not have diabetes mellitus at the baseline visit.
Baseline Covariates and Biomarker Measures Measured at Visit 1
The baseline examination included a home interview, self‐administered questionnaires, and a clinic visit that included blood and urine collections. Other evaluated risk factors included education level, physical activity, and alcohol consumption in the past 12 months. Participants self‐reported the highest level of education attained (less than or greater than a high school diploma). Physical activity was defined as a summary score of the intensity, frequency, and duration of activities encountered daily, which has been previously validated.17
We recorded age, sex, and BMI at the baseline JHS examination. We averaged 2 measures of WC taken at the level of the umbilicus on each participant to determine baseline WC. Hypertension was defined as blood pressure ≥140/90 mm Hg or use of blood pressure–lowering medication. Additional information such as use of diuretics and statins was also collected at each visit. Psychosocial measures including global stress was measured using the Global Perceived Stress Scale, which is an 8‐item measure of global perceptions of stressors associated with ongoing stressful conditions such as employment, relationships, neighborhood, meeting basic needs, and others.18
We asked each participant to fast overnight for at least 12 hours before their clinic visit. We collected blood and urine samples according to the National Committee for Clinical Laboratory Standards as previously reported.19 We measured glucose and insulin concentrations using standard procedures that met the College of American Pathologists accreditation requirements. We used a high‐performance liquid chromatography system (Tosoh Corporation, Tokyo, Japan) to measure glycosylated HbA1c concentrations. We measured total cholesterol from plasma with the cholesterol oxidase method (Roche Diagnostics, Indianapolis, IN) on a Roche COBAS FARA centrifugal analyzer. We measured hs‐CRP using the immunoturbidimetric CRP‐Latex assay (Kamiya Biomedical Company, Seattle, WA) with a Hitachi 911 analyzer (Roche Diagnostics).20 The homeostatic model assessment for insulin resistance (HOMA‐IR) was calculated using the equation HOMA‐IR=(fasting plasma insulin [μU/mL])×(fasting plasma glucose [mmol/L])÷22.5. HOMA‐B, a measure of pancreatic beta cell function, was measured using the equation: HOMA‐B=20×insulin÷fasting plasma glucose−3.5.
We compared baseline characteristics with χ2 tests or 1‐way ANOVA for differences among never, past, and baseline current smokers. We performed multivariable Poisson regression models to estimate the association between smoking status at Visit 1 and incident diabetes mellitus from Visit 1 to Visit 3 yielding incidence rate ratios (IRRs) and 95% confidence intervals (CI). Poisson regression models were utilized to accommodate different follow‐up time intervals among participants between Visit 1 and Visit 3. In the Poisson regression models, the models were offset by log (years), which were calculated as the time between Visit 1 and Visit 3. Covariates included age, sex, BMI, hypertension, total cholesterol, education, physical activity, prevalent cardiovascular disease, diuretic use, statin use, perceived stress, menopause status, and alcohol consumption. Smoking dose (pack‐years) was analyzed as a continuous variable (per 10 pack‐years). We conducted additional secondary analyses adjusting for WC, and adjusting for hs‐CRP. Multivariable linear regression models were used to assess the cross‐sectional associations between smoking status and HOMA‐IR and HOMA‐B at Visit 1. All statistical analyses were performed using SAS 9.4 software (SAS Institute, Cary, NC). All P values were 2‐tailed and P values <0.05 were considered statistically significant.
We studied 2991 participants with follow‐up data at Visit 3. At baseline (Visit 1) we identified 361 current smokers, 502 past smokers, and 2128 never smokers. Table 1 shows the baseline characteristics of the never smokers, past smokers, and current smokers. Compared with never smokers, baseline current and past smokers were more likely to be men (P<0.001). BMI was highest in never smokers and lowest in baseline current smokers (P<0.001). Similarly, WC was lowest in baseline current smokers (P<0.001). Baseline current smokers were younger, less educated, and were more likely to have consumed alcohol in the past year than never smokers and past smokers (all P<0.001). Table 2 shows the baseline characteristics of participants in each analyzed smoking intensity category. High‐intensity cigarette smokers (>20 cigarettes per day) were more likely men, less educated, and more likely to have consumed alcohol in the past 12 months. There were no significant differences in unadjusted fasting glucose levels (P=0.71) or hemoglobin A1c (P=0.67); however, unadjusted plasma insulin levels were lower in high‐intensity smokers (P<0.0001). BMI and WC were also significantly lower in high‐intensity smokers (P<0.0001 and P=0.006, respectively).
Between baseline and Visit 3, 479 participants developed incident diabetes mellitus (mean follow‐up time was 8.0±0.9 years). There were 320 incident diabetes mellitus cases in never smokers (unadjusted incidence rate 15%) compared with 99 in past smokers (unadjusted incidence rate 20%) and 60 in current smokers (unadjusted incidence rate 17%). After multivariable adjustment, smoking status was not significantly associated with risk of incident diabetes mellitus (Table 3). However, further analysis of smoking intensity showed that among baseline current smokers who smoked ≥20 cigarettes daily, there was a significantly higher incidence of diabetes mellitus compared with never smokers (P=0.01), which remained significant after further adjustment for WC (P=0.01). Smoking dose (pack‐years) was also associated with a higher incidence of diabetes mellitus (P=0.02). We further adjusted these models to include hs‐CRP to assess the role of inflammation in the increased risk of diabetes mellitus in smokers. Addition of hs‐CRP only slightly attenuated the risk of incident diabetes mellitus (P=0.03). We also further adjusted for medications including diuretics and statins that have been implicated in hyperglycemia. The findings remained similar after adjusting for diuretic or statin use (IRR 1.73, 95% CI, 1.10–2.73 for high‐intensity smoking and IRR 1.10, 95% CI, 1.01–1.19 for smoking dose). As mental stress has also been implicated in the pathogenesis of diabetes mellitus,21 we additionally adjusted for perceived stress and there were no significant changes in the relationship between high‐intensity smoking (IRR 1.76, 95% CI, 1.12–2.78) or dose (IRR 1.10, 95%, CI 1.01–1.20). The risk of diabetes mellitus has also been reported to increase in postmenopausal women.22 After adjustment for menopause status, the risk for incident diabetes mellitus increased in high‐intensity smokers (IRR 2.01, 95% CI, 1.00–4.06), although not quite statistically significant.
We also assessed insulin resistance (HOMA‐IR) and pancreatic beta cell function (HOMA‐B) to gain mechanistic insight into the relationship between smoking and incident diabetes mellitus. Smoking status and intensity were not significantly associated with HOMA‐IR; however, smoking dose (pack‐years) was associated with a slightly lower HOMA‐IR, suggesting that insulin resistance is not the mechanism of the relationship between smoking and incident diabetes mellitus (Table 4). However, HOMA‐B was lower in current compared with never smokers, and smoking intensity and smoking dose were associated with lower HOMA‐B (Table 5). Therefore, reduced beta cell function is the likely factor leading to our findings.
Our findings from a large cohort of blacks demonstrate high‐intensity cigarette smoking (at least 1 pack per day) and higher pack‐years of smoking were associated with incident diabetes mellitus. Additionally, adjusting for WC and hs‐CRP minimally attenuated the incidence rate.
In 2014, the Surgeon General Report was published on the health consequences of smoking that stated there is sufficient evidence “to infer that cigarette smoking is a cause of diabetes” and that the “risk of developing diabetes is 30% to 40% higher in active smokers than nonsmokers.”23 Several large prospective cohorts have demonstrated increased risk of incident diabetes mellitus in current smokers compared with never smokers. The Nurses' Health Study, a prospective study of over 100 000 women with 24 years of follow‐up, showed a dose‐dependent increased risk of incident type 2 diabetes mellitus in current smokers compared with nonsmokers (relative risk 1.39 for current smokers smoking 1–14 cigarettes per day and relative risk 1.98 for ≥25 cigarettes per day).24 Whereas that study had a very large number of participants, over 93% of the participants were white and they did not analyze specific races separately. In the Insulin Resistance Atherosclerosis Study, current smokers exhibited a significantly higher incidence of diabetes mellitus compared with never smokers (odds ratio 2.66, P=0.001).6 Although many of these studies demonstrated positive and dose‐dependent associations with current cigarette smoking with incident diabetes mellitus, most included very limited numbers of black participants. Other multiethnic cohort studies such as the MESA (Multiethnic Study of Atherosclerosis) did not observe independent associations between cigarette smoking and insulin resistance.5
Given the discrepant published findings and limited data evaluating the link between smoking and incident diabetes mellitus in blacks, our findings have important implications. This is especially important given that almost 18% of blacks are current smokers and blacks have higher rates of diabetes mellitus compared with other ethnic groups.25 Furthermore, many previous studies reporting that smoking was associated with incident diabetes mellitus19 were from a period of time when smoking intensity was higher and composition of cigarettes differed from that of current smokers, including unfiltered cigarettes and higher tar content.26, 27
Inflammation has been proposed as a precursor to the development of type 2 diabetes mellitus.28 Previous studies, including analyses from the JHS, have observed a positive relationship between hs‐CRP (biomarker of inflammation) and incident diabetes mellitus.13, 29 However, this association was largely explained by measures of obesity including BMI and WC. In our study, we further adjusted for hs‐CRP, which only slightly attenuated the incidence rate, suggesting inflammation may play a minor role in the relationship between smoking and incident diabetes mellitus.
We further evaluated relationships of smoking status, intensity, and dose with HOMA‐IR and HOMA‐B, measures of insulin resistance and pancreatic beta cell function, respectively. Only smoking dose was slightly associated with HOMA‐IR, suggesting insulin resistance is not the main contributor to the association of high‐intensity smoking with incident diabetes mellitus. However, HOMA‐B was lower in current smokers compared with never smokers and with increasing smoking dose (pack‐years), suggesting reduced beta cell function is the likely factor leading to our findings.
Our report does have some limitations. First, the JHS is an all‐black cohort, so our findings may not generalize to other ethnic groups. Second, our findings are based on a small number of incident cases of diabetes mellitus, particularly among current smokers. However, to our knowledge, our study is the largest study of the relationship between smoking and incident diabetes mellitus in blacks who have higher rates of diabetes mellitus. Most participants who developed diabetes mellitus were prediabetic (HbA1c 5.7–6.4 mmol/mol). Therefore, current high‐intensity smoking may impose an additional risk in those with a propensity to diabetes mellitus. Cigarette smoking was self‐reported in the JHS; therefore, there is potential for misclassification. Unfortunately, urine cotinine levels were not measured in the participants, which would have provided a more objective exposure level of cigarette smoking. Furthermore, the type of cigarettes that the participants smoked was not available for analysis. We did not account for change in smoking status over follow‐up because of limited availability of longitudinal smoking data in the JHS. As our study was observational, we cannot exclude residual confounding, and cannot infer causality. It is also possible that genetic variation, which we did not assess, may play a role in the impact of smoking on incident diabetes mellitus.30 Last, we cannot adequately account for second‐hand cigarette smoking since such data were not collected at Visit 1.
In conclusion, in a large prospective cohort of blacks, baseline current high‐intensity cigarette smoking and smoking pack years were associated with incident diabetes mellitus in adjusted models. In general, people should be encouraged to stop smoking. Our data suggest that certain high‐risk groups among blacks should be particularly identified for smoking cessation.
Sources of Funding
This research was supported by National Institutes of Health (NIH)/National Heart, Lung, and Blood Institute grant 1P50HL120163‐01. The Jackson Heart Study is supported by contracts HHSN268201300046C, HHSN268201300047C, HHSN268201300048C, HHSN268201300049C, and HHSN268201300050C from the National Heart, Lung, and Blood Institute and the National Institute on Minority Health and Health Disparities. Hall is funded by an NIH/National Institutes of Diabetes and Digestive and Kidney Diseases (NIDDK) grant 1K08DK099415‐01A1 and NIH/National Institute of General Medical Sciences (NIGMS) P20GM104357.
The authors would like to thank the participants of the Jackson Heart Study for their time and participation.
- ↵Mozaffarian D, Benjamin EJ, Go AS, Arnett DK, Blaha MJ, Cushman M, de Ferranti S, Després JP, Fullerton HJ, Howard VJ, Huffman MD, Judd SE, Kissela BM, Lackland DT, Lichtman JH, Lisabeth LD, Liu S, Mackey RH, Matchar DB, McGuire DK, Mohler ER 3rd., Moy CS, Muntner P, Mussolino ME, Nasir K, Neumar RW, Nichol G, Palaniappan L, Pandey DK, Reeves MJ, Rodriguez CJ, Sorlie PD, Stein J, Towfighi A, Turan TN, Virani SS, Willey JZ, Woo D, Yeh RW. American Heart Association Statistics Committee and Stroke Statistics Subcommittee. Heart disease and stroke statistics—2015 update: a report from the American Heart Association. Circulation. 2015;131:e29–e322.
- ↵Okene IS, Miller NH. Cigarette smoking, cardiovascular disease, and stroke. Circulation. 1997;96:3243–3247.
- ↵Keith RJ, Al Rifai M, Carruba C, De Jarnett N, McEvoy JW, Bhatnagar A, Blaha MJ, Defilippis AP. Tobacco use, insulin resistance, and risk of type 2 diabetes: results from the Multi‐Ethnic Study of Atherosclerosis. PLoS One. 2016;11:e0157592.
- ↵Foy CG, Bell RA, Farmer DF, Goff DC, Wagenknecht LE. Smoking and incidence of diabetes among U.S. adults. Diabetes Care. 2005;28:2501–2507.
- ↵Jones MR, Joshu CE, Navas‐Acien A, Platz EA. Racial/ethnic differences in duration of smoking among former smokers in the National Health and Nutrition Examination Surveys. Nicotine Tob Res. 2016;21:ntw326.
- ↵Tuovinen EL, Saarni SE, Männistö S, Borodulin K, Patja K, Kinnunen TH, Kaprio J, Korhonen T. Smoking status and abdominal obesity among normal‐ and overweight/obese adults: population‐based FINRISK study. Prev Med Rep. 2016;4:324–330.
- ↵Effoe VS, Correa A, Chen H, Lacy ME, Bertoni AG. High‐sensitivity C‐reactive protein is associated with incident type 2 diabetes among African Americans: the Jackson Heart Study. Diabetes Care. 2015;38:1694–1700.
- ↵Jackson Heart Study . “JHS Data Access.” Available at: https://www.jacksonheartstudy.org/Research/Study-Data/Data-Access. Accessed November 27, 2017.
- ↵Jee SH, Foong AW, Hur NW, Samet JM. Smoking and risk for diabetes incidence and mortality in Korean men and women. Diabetes Care. 2010;33:2567–2572.
- ↵Payne TJ, Wyatt SB, Mosley TH, Dubbert PM, Guiterrez‐Mohammed ML, Calvin RL, Taylor HA Jr., Williams DR. Sociocultural methods in the Jackson Heart Study: conceptual and descriptive overview. Ethn Dis. 2005;15:S6‐38‐48.
- ↵Fox ER, Benjamin EJ, Sarpong DF, Rotimi CN, Wilson JG, Steffes MW, Chen G, Adeyemo A, Taylor JK, Samdarshi TE, Taylor HA Jr.. Epidemiology, heritability, and genetic linkage of C‐reactive protein in African Americans (from the Jackson Heart Study). Am J Cardiol. 2008;102:835–841.
- ↵U.S. Department of Health and Human Services . The Health Consequences of Smoking—50 Years of Progress: A Report of the Surgeon General. Atlanta, GA: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health; 2014.
- ↵Zhang L, Curhan GC, Hu FB, Rimm EB, Forman JP. Association between passive and active smoking and incident type 2 diabetes in women. Diabetes Care. 2011;34:892–897.
- ↵The changing cigarette. How tobacco smoke causes disease: the biology and behavioral basis for smoking‐attributable disease: a report of the surgeon general. 2010.
- ↵Wisse BE. The inflammatory syndrome: the role of adipose tissue cytokines in metabolic disorders linked to obesity. J Am Soc Nephrol. 2004;15:2792–2800.
- ↵Hu FB, Meigs JB, Rifai N, Manson JE. Inflammatory markers and risk of developing type 2 diabetes in women. Diabetes. 2004;53:693–700.
- ↵Yang J, Zhu Y, Cole SA, Haack K, Zhang Y, Beebe LA, Howard BV, Best LG, Devereux RB, Henderson JA, Henderson P, Lee ET, Zhao J. A gene‐family analysis of 61 genetic variants in the nicotinic acetylcholine receptor genes for insulin resistance and type 2 diabetes in American Indians. Diabetes. 2012;61:1888–1894.