The Association Between Passive Smoking and Type 2 Diabetes

Abstract

The number of people with diabetes has been exponentially increasing. A number of reports in the literature have suggested that exposure to passive smoke may play a key role in the development of diabetes; however, the association has not been jointly summarized yet. In this meta-analysis, 2 databases were searched to identify studies, and the references of these studies were scanned for further studies. Fourteen studies on the relationship between passive smoking and diabetes were included. After all the studies were pooled, the results showed that passive smoking was significantly associated with an increased risk of type 2 diabetes in a random model. The subgroup analysis results were consistent with overall results regardless of type of study design, age, gender, adjustment of dependent variables, area, or study quality. Sensitivity analysis indicated that the overall results were reliable. There was no publication bias observed in the selected studies.

Keywords

passive smoking environmental tobacco smoke type 2 diabetes meta-analysis

Introduction

Type 2 diabetes mellitus (T2DM), one of the most prevalent chronic diseases globally, gives rise to secondary complications, such as cardiovascular disease, renal failure, and blindness.¹ The prevalence of T2DM is anticipated to increase dramatically by 2025; therefore, the identification of risk factors and prevention strategies for the disease are significantly important concerns in public health.² Smoking is the most commonly known risk factor for T2DM. The disease burden attributing to smoking already is enormous with 6 million premature deaths worldwide each year, and is assumed to grow substantially across the century without an end to the pandemic.³

Passive tobacco smoking accounts for an important air pollutant indoors. A recent meta-analysis showed that active tobacco smoking is associated with a 44% increase in the risk of T2DM compared with nonsmokers.⁴ Passive smokers are exposed to similar toxic substances as active smokers, though at different concentration levels.⁵ The views of epidemiological studies on the relation between passive smoking and risk of T2DM among nonsmokers are inconsistent. Recently, Wang et al⁶ conducted a meta-analysis of the association between passive smoking and T2DM based on prospective cohort studies and concluded that there was an association between passive smoking and T2DM. However, we found that the included studies in that meta-analysis were not complete. First, the study by Wang et al⁶ did not include cross-sectional studies and case–control studies with a large number of subjects, which would mean a lot lesser number of subjects. Second, it was strange that their study did not include the study by Houston et al⁷ published in 2006, which should have been searched by their search strategy. There was no reason for the study by Houston et al to be excluded by the inclusion criteria and quality assessment criteria in the study by Wang et al. Moreover, the sample (1452 passive smokers) in the study by Houston et al⁷ was much more than in the studies by Hayashino et al⁸ (690 passive smokers) and Kowall et al⁹ (88 passive smokers). Accordingly, we doubted the credibility of results of the meta-analysis by Wang et al. Therefore, we systematically reviewed and synthesized the existing epidemiological studies to evaluate the association between passive smoking and the risk for T2DM.

Methods

Search Strategy

A comprehensive literature search strategy was carried out in major electronic database (MEDLINE, Web of Science, searched up to August 2013). Text terms used for searching included passive smoking (the exploded index terms secondhand smoking or environmental tobacco smoke [ETS]) in combination with diabetes (the exploded index terms diabetes or diabetes mellitus or prediabetic state or metabolic syndrome X or glucose intolerance or hyperglycemia or glucose metabolism disorders or insulin resistance or glucose tolerance test). We included all sources of passive smoking exposure (parents, household, other family members) measured by either self-report or cotinine measure. In addition, the reference lists of retrieved studies to identify further studies were also searched.

Study Selection

We limited our review to observational studies that (a) involved human subjects, (b) included the association between passive smoking and risk of T2DM among nonsmokers, and (c) were published in peer-reviewed scientific journals in English language only. Titles and abstracts identified from the searches were checked for eligibility independently by 2 authors (either BZ and XW, or XW and QZ). Studies, which were definitely not to be relevant, were excluded at each stage. The full texts of potentially eligible articles were searched and also checked for their eligibility independently by 2 authors. Disagreements were resolved through discussion with a third author.

The studies had to meet the following criteria to be eligible for inclusion in our meta-analysis: (a) The outcome is T2DM. Both of the diagnostic criteria for diabetes of the American Diabetes Association (ADA) and the World Health Organization (WHO) had been widely used in epidemiological surveys, so we included studies that used the criteria of ADA or WHO to diagnose diabetes. (b) The exposure of passive smoking was ascertained by either self-report (participants who were exposed to passive smoke but did not actively smoke were asked by questionnaire) or cotinine measure (participants were exposed to passive smoke and also had low cotinine concentrations in serum [1-15 ng/mL]¹⁰). (c) The study reported calculated odds ratio (OR), relative risk (RR), hazard ratio (HR), or reported data.

Seventy-five studies from MEDLINE and 125 studies from Web of Science were identified by the systematic search on electronic databases. After looking through the titles and abstracts, we excluded the duplicated studies, reviews, nondiabetic studies, and experimental studies. After reading the full articles, 14 eligible studies were selected into our systematic review and meta-analysis (Figure 1).

Data Extraction

Data from the included studies relating to source, area, gender, diabetes measure, exposure examined, year, and results were extracted independently by 2 authors (either BZ and XW, or XW and QZ), and we extracted each reported OR, RR, HR, or incidence density ratios for the risk of developing diabetes or other glucose metabolism irregularities for passive smokers compared with nonsmokers. Both unadjusted and adjusted estimates (OR, RR, and HR) were extracted from each study. To reduce effects of confounding factors on the results, we preferred to extract and analyze the adjusted estimates to unadjusted estimates. Confounders were identified by reading the articles. The studies, which solely reported categories of passive smoking, were classified into the highest category for the analysis. If one study presented several estimates adjusted for different combinations of potential confounders, the one that adjusted for the greatest number of potential confounders were selected.

Quality Assessment

Two authors (BZ and XW) independently assessed the included studies with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement.¹¹ Seven aspects were selected for all types of studies: (a) Settings: describe the settings, location, and relevant dates, including periods of exposure, follow-up, and data collection; (b) Participants: give the eligibility criteria, and the sources and methods of selection of participants; (c) Variables: clearly define all outcomes, exposures, predictors, potential confounders, and effect modifiers. Give diagnostic criteria, if applicable; (d) Measurement: For each variable of interest, clarify sources of data and details of methods of assessment (measurement). Describe comparability of assessment methods if there is more than one group; (e) Statistical methods: Describe all statistical methods, including those used to control for confounding; (f) Descriptive data: depict the characteristics of study participants (eg, demographic, clinical, social) and information on exposures and potential confounders; (g) Main results: Elaborate unadjusted estimates and, if applicable, confounder-adjusted estimates and their precision (eg, 95% confidence intervals [95% CIs]). Clearly make sure which confounders were adjusted for and why they were included. The study that met the 7 items was regarded as “high quality,” the study that met less than 5 items was regarded as “low quality,” and otherwise, the study was regarded as “moderate quality.”

Statistical Analysis

Meta-analysis was performed by STATA V.12 (StataCorp, College Station, TX). We evaluated heterogeneity across studies using the Cochrane Q statistic. Heterogeneity was quantified using recognized methods (I²), this statistic yields results ranged from 0% to 100% (I² = 0% to 25%, no heterogeneity; I² = 25% to 50%, moderate heterogeneity; I² = 50% to 75%, large heterogeneity; I² = 75% to 100%, extreme heterogeneity).¹² If heterogeneity existed, the random-effects model was used; otherwise, the fixed model was used.

We also conducted subgroup analysis by type of study design (cohort studies, case–control studies, and cross sectional studies), age (<50 and ≥50 years old), gender (both, men, and women), adjustment of dependent variables (yes, no), area (United States, Europe, and Asia), and study quality (high, moderate, and low). In addition, we investigated the influence of a single study on the overall risk estimate by removing each study in each turn, to test the effect for the overall results.

Publication bias was assessed by the Begg and the Egger tests and visual inspection of funnel plots. Statistical tests with a 2-sided P value <.05 were considered significantly different.

Results

Study Characteristics

Fourteen published studies^7-9,13-23 were identified and included for this review and meta-analysis: 8 studies^{7-9,13,14,16,18,20,22} specifically investigated the relationship between passive smoking and risk of T2DM among nonsmokers(5 studies^7-9,18,22 were prospective cohort design, 2 studies^13,20 were cross-sectional design, and 1 study¹⁶ was a case-control design). The remaining 6 studies^{14,15,17,19,21,23} investigated the relationship between passive smoking and other diseases (such as arterial atherosclerosis, tuberculosis, and acute coronary syndrome, etc), including the data between passive smoking and diabetes. Five of the studies were conducted in the United States, 5 in European countries, 4 in Asian countries. Eleven studies included both men and women, while the remaining 3 studies included only women. All included studies are summarized in Table 1.

Table 1.

Studies Included in the Systematic Review.

Source	Type of Study Design	Area	Gender	No. of Cases	Age in Years (Mean or Median)	Diabetes Measure	Exposure Examined	Year	Adjustments for Confounders^a
Vardavas et al¹³	Cross-sectional study	Mediterranean Islands	Both	263	>65	Biologically screened	Home or work	2009	Age, sex, hypertension, hypercholesterolemia, family history of diabetes, physical activity status, BMI, years of school, MedDietScore
Seki et al¹⁴	Cross-sectional study	Japan	Both	35	54	Patient reported	Home or work	2010	No
Henkin et al¹⁵	Cross-sectional study	USA	Both	191	53	Biologically screened	Home or work	1999	No
Hayashino et al⁸	Cohort study	Japan	Both	—	39.3	Mixed methods	Home or work	2008	Age, sex, BMI, physical activity, alcohol, family history of diabetes, hypertension, health promotion intervention, frequency of sweetened beverage intake, frequency of vegetable intake
Hargrave et al¹⁶	Case–control study	Northern Ireland	Both	100	42.5	Biologically screened	—	1999	No
Panagiotakos et al¹⁷	Case–control study	Greece	Both	83	47	Biologically screened	Home or work	2004	No
Houston et al⁷	Cohort study	USA	Both	1452	24	Biologically screened	—	2006	Age, sex, ethnicity, education, income, physical activity, daily total caloric intake, daily saturated fat caloric intake, alcohol consumption, systolic blood pressure, triglycerides, physician visits, health insurance, WHR, serum insulin, C-reactive protein
Ko et al¹⁸	Cohort study	Korea	Both	465	50.5	Mixed methods	Home or work	2011	Age, sex, residential area, education level, alcohol consumption, waist circumference, regular exercise, hypertension history, total cholesterol, HOMA-IR, baseline glucose tolerance status
Leung et al¹⁹	Cross-sectional study	China	Both	2209	68.6	Biologically screened	Home	2010	Age, Cantonese speaking, education, housing, on public assistance, alcohol use
Kowall et al⁹	Cohort study	Germany	Both	70	60	Biologically screened	Home or work	2010	Age, sex, parental diabetes, socioeconomic status, alcohol intake, physical activity, intake of meat and sausage, intake of salad and vegetables, intake of whole-grain bread, and coffee consumption, waist circumference, blood pressure, hypertriglyceridemia, HDL-cholesterol, log insulin, log adiponectin
Forastiere et al²⁰	Cross-sectional study	Italy	Women	66	—	Biologically screened	Home	2000	Center, age, center × age, and education
Howard et al²¹	Cross-sectional study	USA	Both	405	54	Mixed methods	Home or work	1998	No
Zhang et al²²	Cohort study	USA	Women	5932	47	Patient reported	—	2011	Age, race, BMI, physical activity, husband’s education, family history of diabetes, total energy intake, and intake of alcohol, magnesium, calcium, vitamin D, total trans fat, fiber from cereal, caffeine, total fat, and saturated fat
Kawachi et al²³	Cross-sectional study	USA	Women	815	60	Registry consultation	Home or work	1997	No

Abbreviations: HOMA-IR, homeostasis model assessment of insulin resistance; BMI, body mass index; WHR, waist to hip ratio; HDL, high-density lipoprotein.

The confounders were adjusted by multivariable analysis.

Quality Assessment

Table 2 shows the result of the quality assessment of the included studies. It shows that 6 studies were of high quality, 7 items were described in detail. Six studies were of moderate quality, as they did not give confounder-adjustment estimates. One study was of low quality, because 2 items were not described and 1 item was described only partly.

Table 2.

Quality Assessment for Included Studies.

	Setting	Participants	Variables	Measurement	Statistical Methods	Descriptive Data	Main Results
Vardavas et al¹³	•	•	•	◎	°	°	•
Seki et al¹⁴	•	•	•	•	•	•	◎
Henkin et al¹⁵	•	•	•	•	•	•	◎
Hayashino et al⁸	•	•	•	•	•	•	•
Hargrave et al¹⁶	•	•	•	•	•	◎	◎
Panagiotakos et al¹⁷	•	•	•	•	•	•	◎
Houston et al⁷	•	•	•	•	•	•	•
Ko et al¹⁸	•	•	•	•	•	•	•
Leung et al¹⁹	•	•	•	•	•	•	•
Kowall et al⁹	•	•	•	•	•	•	•
Forastiere et al²⁰	•	•	•	•	•	•	•
Howard et al²¹	•	•	•	•	•	•	◎
Zhang et al²²	•	•	•	•	•	•	•
Kawachi et al²³	•	•	•	•	•	•	◎

•, the item was described in detail; ◎, the item was described partly; °, the item was not described.

Overall Risk of Diabetes for Passive Smoker

Table 3 shows that passive smoking was significantly associated with an increased risk of T2DM in overall studies when using random-effects model (RR = 1.26, 95% CI = 1.15-1.37). The overall heterogeneity of the studies was moderate (I² = 35.4%). Figure 2 shows the association between passive smoking and T2DM risk using a random-effects model.

Table 3.

Overall and Subgroup Analyses for the Risk of Diabetes in Passive Smokers.

		Pooled RR (95% CI)
Factor	No. of Studies	Fixed-Effects Model	Random-Effects Model	I² (%)
Overall studies	14	1.22 (1.15-1.30)	1.26 (1.15-1.37)	35.40
Study design
Cross-sectional studies	7	1.19 (1.10-1.28)	1.23 (1.09-1.38)	38.40
Cohort studies	5	1.30(1.17-1.44)	1.34 (1.15-1.56)	36.70
Case–control studies	2	1.16 (0.88-1.53)	1.13 (0.75-1.71)	53.30
Age in years^a
<50	5	1.22 (1.09-1.37)	1.24 (1.07-1.44)	27.20
≥50	8	1.23 (1.15-1.32)	1.29 (1.15-1.46)	47.20
Gender
Both	11	1.22 (1.14-1.31)	1.28 (1.15-1.44)	39.70
Female	3	1.23 (1.10-1.37)	1.22 (1.15-1.37)	43.30
Male	0	—	—	—
Adjustment of dependent variables
Yes	8	1.20 (1.12-1.29)	1.27 (1.12-1.44)	47.80
No	6	1.27 (1.14-1.41)	1.26 (1.11-1.44)	18.70
Area
USA	5	1.25 (1.15-1.36)	1.26 (1.14-1.39)	21.90
Europe	5	1.24 (1.01-1.52)	1.25 (0.94-1.67)	45.00
Asia	4	1.19 (1.09-1.30)	1.29 (1.05-1.58)	57.70
Quality
High	7	1.19 (1.11-1.29)	1.25 (1.10-1.41)	48.80
Moderate	6	1.27 (1.14-1.41)	1.26 (1.11-1.44)	18.70
Low	1	1.60 (1.05-2.51)	1.60 (1.05-2.51)	—

Abbreviations: RR, relative risk; CI, confidence interval.

There was no the mean or median of age in 1 study in the subgroup.

Subgroup Analyses for Factors

As shown in Table 3, according to the type of study design, the pooled estimate of cohort studies was 1.34 (95% CI = 1.15-1.56) and the pooled estimate of cross-sectional studies was 1.23 (95% CI = 1.09-1.38); there was no significant differences in that of case–control studies (RR = 1.13, 95% CI = 0.75-1.71).

In terms of age, because in 1 study²⁰ the mean or median age was not calculated, only 13 studies were analyzed. The risk of T2DM in the subgroups aged older than 50 years was 1.29 (95% CI = 1.15-1.44), and the risk of T2DM within 50 years of age was 1.24 (95% CI = 1.07-1.44). In a subgroup analysis of studies with adjustment of dependent variables, the associations between passive smoking and risk of T2DM were consistent with the overall results. In the included studies, there were 5 studies conducted in the United States, 5 in Europe, and 4 in Asia. The risk of T2DM in the United States (RR = 1.25, 95% CI = 1.15-1.36) and Europe (RR = 1.25, 95% CI = 0.94-1.67) was slightly lower than that in Asia (RR = 1.29, 95% CI = 1.05-1.58). Although the risk in Europe failed to reach statistical significance, it showed similarity with other 2 regions. When we grouped studies by quality, all subgroups showed a higher risk of T2DM in passive smokers.

Sensitivity Analysis

Table 4 shows the pooled RRs and 95% CIs of sensitivity analysis by removing 1 study in each turn; the results indicated that the overall result was reliable.

Table 4.

Sensitivity Analysis by Removing Each Study in Each Turn.

		95% Confidence Interval
Study Removed	Relative Risk	Lower Limit	Upper Limit
Vardavas et al¹³	1.25	1.14	1.36
Seki et al¹⁴	1.26	1.15	1.38
Henkin et al¹⁵	1.24	1.14	1.36
Hayashino et al⁸	1.25	1.14	1.36
Hargrave et al¹⁶	1.25	1.15	1.37
Panagiotakos et al¹⁷	1.27	1.17	1.38
Houston et al⁷	1.25	1.14	1.37
Ko et al¹⁸	1.24	1.13	1.36
Leung et al¹⁹	1.29	1.18	1.41
Kowall et al⁹	1.25	1.15	1.36
Forastiere et al²⁰	1.27	1.16	1.38
Howard et al²¹	1.28	1.16	1.41
Zhang et al²²	1.28	1.16	1.41
Kawachi et al²³	1.24	1.13	1.36

Assessment of Publication Bias

Figure 3 shows that neither the Begg nor the Egger test provided evidence for statistically significant publication bias for any of the overall meta-analysis (Begg, P = .443; Egger, P = .113), but the funnel plot suggested slight asymmetry.

Discussion

In this meta-analysis, the studies were included after comprehensive literature searches, and of the studies in their reference lists were searched as well. The results showed that there was a significantly positive association between passive smoking and the risk of T2DM. Passive smoking was associated with a 26% increased risk of T2DM in our meta-analysis, which was slightly lower in RR than active smoking.⁴ This finding was consistently observed in subgroup analysis by the type of study design, age, gender, adjustment of dependent variables, area, or study quality. Sensitivity analysis indicated that the overall result was reliable.

In all included studies, information on passive smoking was mostly collected by questionnaires, which relied on self-reports of passive smoking. Ideally, the exposure should be measured in terms of the biological dose of the contaminants or its metabolites received by the target tissue. Current studies only collected the intensity and/or the duration of the exposure to smoke using questionnaires or interviews, which were unable to accurately estimate the dose of tobacco received by individuals.²⁴ In addition, intensity of exposure to passive smoke would be influenced by not only the frequency and duration of smoke exposure but also the number of smokers, the number of cigarettes smoked, the volume of space, and ventilation equipment. A wide variation was noted in passive smoke concentration depending on locations, including offices, restaurants, bars, and residences.²⁵ Therefore, this well explained why the results of the studies were inconsistent between passive smoking and T2DM.

Additionally, the pooled RR of passive smoking was lower than that of active smoking in the risk of diabetes. Although passive smoking might exert a smaller effect than active smoking at an individual level,¹⁹ passive smoking might exert a much bigger impact than active smoking at the population level, because everyone breathing the same air could be affected, both smokers and never-smokers. For example, passive smoking contributed to 13.7% of active tuberculosis, while active, current, and previous smoking only contributed to 8.6% of active tuberculosis in a cohort study.²⁶ In short, the risk of passive smoking could be underestimated.

It is acknowledged that there exist several limitations to what conclusions can be drawn from the meta-analysis. First, our findings were based on the results of observational studies (7 cross-sectional studies, 5 cohort studies, and 2 case-control studies); this might make the results unreliable. Therefore, we performed a stratified analysis to explore it and found out that the pooled estimates of cohort studies (RR = 1.34, 95% CI = 1.15-1.56), cross-sectional studies (RR = 1.23, 95% CI = 1.09-1.38), and case–control studies (RR = 1.13, 95% CI = 0.75-1.71) were consistent with the overall results. It should be noted that the pooled RR in cohort studies, which have more reliability than others, was the highest, indicating that other types of studies may reduce the risk of T2DM on passive smoking. Second, all included studies were observational studies, which did not exclude the existence of unidentified confounders. Eight studies in the meta-analysis adjusted results for confounders, which identified 4 to 15 confounders. The results in all included studies were not entirely adjusted dependent variables, so the 95% CI was wider for nonadjustment than adjustment. To address the problem, we performed a stratified analysis and found that the pooled estimate of studies in adjustment of dependent variables was 1.27 (95% CI = 1.12-1.44); the associations between passive smoking and risk of T2DM were consistent with the overall results. Third, there is inconsistency in the identification criteria of type 2 diabetes in included studies. Although some studies took some measures to reduce the misclassification bias for case identification, self-reporting may cause misclassification of the outcome. Finally, publication bias could be of concern in meta-analysis. Nevertheless, we found no evidence of publication bias.

The International Agency for Research on Cancer concluded that passive smoke is carcinogenic to humans as group I.²⁷ The adverse effects of passive smoking have attracted considerable attention in the world. There have been some epidemiological studies on the relationship between passive smoking and chronic diseases, such as heart disease,^17,23 cardiovascular disease,²⁸ cancer,^24,27 and diabetes^9,13,20,22; our study focused on the relation between passive smoking and diabetes. Despite heterogeneity among the included studies, we found that the results from this meta-analysis with current data indicated a statistically significant positive association between passive smoking and risk of type 2 diabetes. These findings suggest that passive smoking may increase the prevalence of diabetes and that measures of diabetes control should include reducing tobacco exposure, especially among those at high risk for diabetes.

Conclusion

Passive smoking is associated with an increased risk of type 2 diabetes. Therefore, interventions to prevent exposure to passive smoke remain an urgent priority.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the National Natural Science Foundation of China (Grant No. 81072243).

References

Van Dieren

Beulens

van der Schouw

Grobbee

Neal

. The global burden of diabetes and its complications: an emerging pandemic. Eur J Cardiovasc Prev Rehabil. 2010;17(suppl 1):S3-S8.

Heidari

Myers

. Brief update on the burden of diabetes in south carolina. Am J Med Sci. 2013;345:302-306.

Wipfli

Samet

. Global economic and health benefits of tobacco control: part 1. Clin Pharmacol Ther. 2009;86:263-271.

Willi

Bodenmann

Ghali

Faris

Cornuz

. Active smoking and the risk of type 2 diabetes. JAMA. 2007;298:2654-2664.

Chan-Yeung

Dimich-Ward

. Respiratory health effects of exposure to environmental tobacco smoke. Respirology (Carlton, Vic.). 2003;8:131-139.

Wang

Liu

Deng

. Passive smoking and risk of type 2 diabetes: a meta-analysis of prospective cohort studies. PLoS One. 2013;8:e69915.

Houston

Person

Pletcher

Liu

Iribarren

Kiefe

. Active and passive smoking and development of glucose intolerance among young adults in a prospective cohort: CARDIA study. BMJ. 2006;332:1064-1069.

Hayashino

Fukuhara

Okamura

; HIPOP-OHP Research Group. A prospective study of passive smoking and risk of diabetes in a cohort of workers: the High-Risk and Population Strategy for Occupational Health Promotion (HIPOP-OHP) study. Diabetes Care. 2008;31:732-734.

Kowall

Rathmann

Strassburger

. Association of passive and active smoking with incident type 2 diabetes mellitus in the elderly population: The KORA S4/F4 cohort study. Eur J Epidemiol. 2010;25:393-402.

10.

Vineis

Alavanja

Buffler

. Tobacco and cancer: recent epidemiological evidence. J Natl Cancer Inst. 2004;96:99-106.

11.

Little

Higgins

Ioannidis

. Strengthening the REporting of Genetic Association studies (STREGA)—an extension of the STROBE statement. Eur J Clin Invest. 2009;39:247-266.

12.

Higgins

Thompson

. Quantifying heterogeneity in a meta-analysis. Stat Med. 2002;21:1539-1558.

13.

Vardavas

Lionis

Polychronopoulos

. The role of second-hand smoking on the prevalence of type 2 diabetes mellitus in elderly men and women living in Mediterranean islands: The MEDIS study. Diabet Med. 2010;27:242-243.

14.

Seki

Inoue

Ohkubo

. Association of environmental tobacco smoke exposure with elevated home blood pressure in Japanese women: The Ohasama study. J Hypertens. 2010;28:1814-1820.

15.

Henkin

Zaccaro

Haffner

. Cigarette smoking, environmental tobacco smoke exposure and insulin sensitivity: the Insulin Resistance Atherosclerosis Study. Ann Epidemiol. 1999;9:290-296.

16.

Hargrave

McMaster

O’Hare

Carson

. Tobacco smoke exposure in children and adolescents with diabetes mellitus. Diabet Med. 1999;16:31-34.

17.

Panagiotakos

Chrysohoou

Pitsavos

. The association between secondhand smoke and the risk of developing acute coronary syndromes, among non-smokers, under the presence of several cardiovascular risk factors: The CARDIO2000 case-control study. BMC Public Health. 2002;2:9.

18.

Min

Ahn

. A prospective study investigating the association between environmental tobacco smoke exposure and the incidence of type 2 diabetes in never smokers. Ann Epidemiol. 2011;21:42-47.

19.

Leung

Lam

. Passive smoking and tuberculosis. Arch Intern Med. 2010;170:287-292.

20.

Forastiere

Mallone

Presti

. Characteristics of nonsmoking women exposed to spouses who smoke: epidemiologic study on environment and health in women from four Italian areas. Environ Health Perspect. 2000;108:1171-1177.

21.

Howard

Wagenknecht

Burke

. Cigarette smoking and progression of atherosclerosis: the Atherosclerosis Risk in Communities (ARIC) study. JAMA. 1998;279:119-124.

22.

Zhang

Curhan

Rimm

Forman

. Association between passive and active smoking and incident type 2 diabetes in women. Diabetes Care. 2011;34:892-897.

23.

Kawachi

Colditz

Speizer

. A prospective study of passive smoking and coronary heart disease. Circulation. 1997;95:2374-2379.

24.

Zhou

Wellenius

Michaud

. Environmental tobacco smoke and the risk of pancreatic cancer among non-smokers: a meta-analysis. Occup Environ Med. 2012;69:853-857.

25.

Siegel

. Involuntary smoking in the restaurant workplace: a review of employee exposure and health effects. JAMA. 1993;270:490-493.

26.

Leung

Lam

. Smoking and tuberculosis among the elderly in Hong Kong. Am J Respir Crit Care Med. 2004;170:1027-1033.

27.

Wagenknecht

Burke

Perkins

Haley

Friedman

. Misclassification of smoking status in the CARDIA study: a comparison of self-report with serum cotinine levels. Am J Public Health. 1992;82:33-36.

28.

Dunbar

Gotsis

Frishman

. Second-hand tobacco smoke and cardiovascular disease risk: an epidemiological review. Cardiol Rev. 2013;21:94-100.