SMOKING

182 Am J Prev Med 2020;58(2):182−190 © 20

RESEARCH ARTICLE

From the 1Cente sity of California Family Compreh cisco, San Franci University of Cali vascular Research Francisco, Califor ies, University of

Address corr Tobacco Control Francisco, 530 Pa mail: stanton.glan

0749-3797/$36 https://doi.org

19 American Journal of Preven

Association of E-Cigarette Use With Respiratory

Disease Among Adults: A Longitudinal Analysis

Dharma N. Bhatta, PhD, MPH,1,2 Stanton A. Glantz, PhD1,2,3,4,5

Introduction: E-cigarettes deliver an aerosol of nicotine by heating a liquid and are promoted as an alternative to combustible tobacco. This study determines the longitudinal associations between e-cigarette use and respiratory disease controlling for combustible tobacco use.

Methods: This was a longitudinal analysis of the adult Population Assessment of Tobacco and Health Waves 1, 2, and 3. Multivariable logistic regression was performed to determine the associa- tions between e-cigarette use and respiratory disease, controlling for combustible tobacco smoking, demographic, and clinical variables. Data were collected in 2013−2016 and analyzed in 2018−2019.

Results: Among people who did not report respiratory disease (chronic obstructive pulmonary disease, chronic bronchitis, emphysema, or asthma) at Wave 1, the longitudinal analysis revealed statistically significant associations between former e-cigarette use (AOR=1.31, 95% CI=1.07, 1.60) and current e-cigarette use (AOR=1.29, 95% CI=1.03, 1.61) at Wave 1 and having incident respiratory disease at Waves 2 or 3, controlling for combustible tobacco smoking, demographic, and clinical varia- bles. Current combustible tobacco smoking (AOR=2.56, 95% CI=1.92, 3.41) was also significantly associated with having respiratory disease at Waves 2 or 3. Odds of developing respiratory disease for a current dual user (e-cigarette and all combustible tobacco) were 3.30 compared with a never smoker who never used e-cigarettes. Analysis controlling for cigarette smoking alone yielded similar results.

Conclusions: Use of e-cigarettes is an independent risk factor for respiratory disease in addition to combustible tobacco smoking. Dual use, the most common use pattern, is riskier than using either product alone. Am J Prev Med 2020;58(2):182−190. © 2019 American Journal of Preventive Medicine. Published by Elsevier Inc. All rights reserved.

INTRODUCTION

r for Tobacco Control Research and Education, Univer- , San Francisco, San Francisco, California; 2Helen Diller ensive Cancer Center, University of California, San Fran- sco, California; 3Department of Medicine (Cardiology), fornia, San Francisco, San Francisco, California; 4Cardio- Institute, University of California, San Francisco, San nia; and 5Philip R. Lee Institute for Health Policy Stud- California, San Francisco, San Francisco, California espondence to: Stanton A. Glantz, PhD, Center for Research and Education, University of California San rnassus Avenue, Suite 366, San Francisco CA 94143. E- [email protected]. .00 /10.1016/j.amepre.2019.07.028

R espiratory diseases are leading causes of morbid- ity and mortality in the U.S.1,2 Smoking is a major cause3 and, like combustible tobacco

products, e-cigarettes expose users to nicotine, ultrafine particles, and other toxicants.4 Some pulmonary toxi- cants are in e-cigarette aerosol at higher levels than com- busted cigarettes, including propylene glycol,5 diacetyl6,7

(butter flavor), cinnamaldehyde8 (cinnamon), benzalde- hyde (cherry), and metals.9,10

Animal studies found that e-cigarettes increase pulmo- nary inflammation and oxidative stress while inhibiting the immune response.11 Repeated exposure to acrolein produced by heating propylene glycol and glycerin in e-liquids causes chronic pulmonary inflammation, reduc- tion of host defense, neutrophil recruitment and

activation, mucus hypersecretion, and protease-mediated lung tissue damage, which are linked to development of chronic obstructive pulmonary disease12 (COPD). Mice exposed to nicotine e-cigarette aerosol exhibit increased

tive Medicine. Published by Elsevier Inc. All rights reserved.

Bhatta and Glantz / Am J Prev Med 2020;58(2):182−190 183

airway and alveolar cell death and airspace enlargement similar to COPD,13 and rats suffer emphysematous air- space enlargement and loss of lung vascular elements.14

E-cigarette exposure depresses pulmonary immune defenses against viral and bacterial infection in mice.15

Inhalation of nicotine e-cigarette aerosol disrupts airway barrier function and induces systemic inflammation in mice.16 Consistent with these experimental results, people who use e-cigarettes experience decreased expression of immune-related genes in their nasal cavities, with more genes suppressed than among cigarette smokers, indicat- ing immune suppression in the nasal mucosa.17 E-cigarette use upregulates expression of platelet-activating factor receptor in users’ nasal epithelial cells,18 an important molecule involved in the ability of Streptococcus pneumo- niae, the leading cause of bacterial pneumonia, to attach to cells that it infects. E-cigarette users exhibit significant increases in aldehyde detoxification− and oxidative stress−related proteins associated with cigarette smoke, providing additional evidence that e-cigarettes may adversely affect the profile of innate defense proteins in airway secretions similar to that observed among cigarette smokers.19 Epithelial cells from human lung biopsy sam- ples reveal that about 300 proteins are differentially expressed in smoker and e-cigarette user airways, with only 78 proteins commonly altered in both groups, sug- gesting that the propylene glycol/vegetable glycerin carrier used in e-cigarettes might explain the differences.20

Consistent with the biology, cross-sectional studies found associations between e-cigarettes and respiratory disease among children21−23 and adults (Perez et al., 2018. E-cigarette use is associated with emphysema, chronic bronchitis and COPD. In: American Thoracic Society 2018 International Conference).24 A longitudinal study of individuals with COPD found that e-cigarette use was associated with chronic bronchitis and COPD exacerbations and more rapid decline in lung function, adjusting for tobacco smoking.25

This paper uses the first 3 waves of the public use data files for the Population Assessment of Tobacco and Health (PATH) to determine the longitudinal associa- tion between e-cigarette use and respiratory diseases, controlling for combustible tobacco use and other risk factors in a large representative sample of U.S. adults.

METHODS Data were collected in 2013−2016 and analyzed in 2018−2019.

Study Population This study used the adult (aged ≥18 years) sample in PATH Waves 1 (September 2013 to December 2014), 2 (October 2014 to October 2015), and 3 (October 2015 to October 2016), a nationally representative, population-based, longitudinal study

February 2020

(Appendix Figure 1, available online). The weighted response rate at Wave 1 household screener was 54.0%; among screened house- holds, the overall weighted response rate at Wave 1 adult inter- view was 74.0%. The weighted adult retention rates at Waves 2 and 3 were 83.2% and 78.4%, respectively. The University of California San Francisco Committee on Human Research ruled this study exempt.

Measures Lung or respiratory disease at Wave 1 was assessed with the ques- tion: Has a doctor or other health professional ever told you that you had any of the following lung or respiratory conditions? (yes or no): COPD, chronic bronchitis, emphysema, and asthma. Respond- ents who answered yes to any of these questions were coded as having lung or respiratory disease at Wave 1.

Lung or respiratory disease at Waves 2 and 3 was assessed with the question: In the past 12 months, has a doctor, nurse, or other health professional told you that you had any of the following lung or respiratory conditions? (yes or no): COPD, chronic bronchitis, emphysema, and asthma. Respondents who answered yes to any of these questions were coded as having lung or respiratory disease at Wave 2 or 3.

Respondents who ever used an e-cigarette, ever used fairly reg- ularly, and currently used every day or some days were considered current users. Respondents who reported that they ever used e- cigarettes but do not currently use e-cigarettes were considered former users. Respondents who reported that they have never used e-cigarettes, even once or twice, were considered never users.

Respondents who currently smoked cigarettes, traditional cigars, filtered cigars, cigarillos, pipe tobacco, or hookah every day or some days (regardless of whether they have smoked 100 ciga- rettes in their lifetime) were considered current combustible tobacco smokers. Respondents who ever smoked and currently do not smoke at all were classified as former smokers. Respondents who reported that they have never smoked, even 1 or 2 puffs, were classified as never smokers.

The same definitions were used to define conventional cigarette smoking status.

Demographic variables assessed at Wave 1 were age, BMI, sex (male or female), race/ethnicity (white, black, and other), and poverty level (below or above 100% of the poverty line).

InWave 1, respondents who answered yes toHas a doctor, nurse, or other health professional ever told you that you had high blood pressure? were coded as having high blood pressure. Respondents who answered yes toHas a doctor, nurse, or other health professional ever told you that you had high cholesterol? were coded as having high cholesterol. Respondents who answered yes to Has a doctor, nurse, or other health professional ever told you that you had diabe- tes, sugar diabetes, high blood sugar, or borderline diabetes? were coded as having diabetes mellitus.

Statistical Analysis Logistic regression was used to quantify cross-sectional associa- tion between e-cigarette use (former and current) and respiratory disease at Wave 1, controlling for combustible tobacco smoking (former and current), age, BMI, sex, poverty level, race/ethnicity, and clinical variables. The reference condition was people who had never used e-cigarettes or smoked combusted tobacco prod- ucts (cigarettes in the subsidiary analysis).

Table 1. Demographic, Clinical, and Tobacco Use Variables at Wave 1 Baseline (n=32,320)

Variables Weighted %

Respiratory disease

Yes 15.1

No 84.9

Tobacco use

E-cigarette user

Never 82.3

Former 12.2

Current 5.5

Combustible tobacco smoker

Never 28.6

Former 45.4

Current 26.0

Cigarette smoker

Never 33.2

Former 45.4

Current 21.4

Demographic

184 Bhatta and Glantz / Am J Prev Med 2020;58(2):182−190

Among respondents who did not report any respiratory disease at Wave 1, logistic regression was used to quantify the longitudi- nal association between e-cigarette use at Wave 1 and incident respiratory disease at either Wave 2 or Wave 3 combined, control- ling for combustible tobacco smoking (former and current), age, BMI, sex, poverty level, race/ethnicity, and clinical variables at Wave 1. Waves 2 and 3 were combined to increase the number of events and the power of the study, essentially treating the study as a 2-year longitudinal follow up from baseline when e-cigarette use was assessed.

A separate analysis was performed on the effect of e-cigarette use on respiratory disease after controlling for cigarette smoking only, demographic, and clinical variables.

The PATH-provided different weights for the cross-sectional and follow up data sets were used as specified in the PATH Study user guide.26 Survey package, version 3.33-2 in R was used for sta- tistical analyses accounting for the complex survey design.

There are very little missing data in PATH. The number of dropped cases was only 1,028 (respiratory disease, n=127; e-ciga- rette users, n=42; any combustible tobacco smokers, n=774; con- ventional cigarette smokers, n=85), 5.3% of the sample. Given the very low level of missing data, list-wise deletion was used.

Age in years

18−24 13.1

25−34 17.7

35−44 16.5

45−54 17.9

55−64 16.6

65−74 11.1

75 and above 7.1

BMI (§SD), kg/m2 28.00 (§6.8)

Sex

Male 48.1

Female 51.9

Poverty level/income

Below poverty (<100% of poverty guideline)

25.2

At or above poverty (≥100% of poverty guideline)

74.8

Race/ethnicity

White 77.9

Black 12.3

Other 9.8

High blood pressure

Yes 27.8

No 72.2

High cholesterol

Yes 23.0

No 77.0

Diabetes mellitus

Yes 14.0

No 86.0

RESULTS

Table 1 shows baseline descriptive statistics and Appen- dix Table 1, available online, shows the relationships between e-cigarette use and combusted tobacco and cig- arette smoking. A total of 5,466 (15.1%) adults reported that they had respiratory disease at baseline. Table 2 shows the descriptive statistics stratified by respiratory disease at Wave 1 and combined Waves 2 and 3. Appen- dix Table 2, available online, reports detailed informa- tion by specific diagnosis. Among people who did not report respiratory dis-

ease at Wave 1, tobacco users who reported new respi- ratory disease at Waves 2 or 3 tended to be more addicted, as measured by shorter time to first tobacco product use and frequency of tobacco product use (Appendix Table 3, available online). There were no differences in use of flavored tobacco products (Appendix Table 4, available online). Table 3 (left columns) shows the cross-sectional associ-

ations between e-cigarette use and having had respiratory disease at Wave 1 adjusting for combustible tobacco smoking, demographic, and clinical variables. The risk of having had respiratory disease was significantly associated with former e-cigarette use (AOR=1.34, 95% CI=1.23, 1.46) and current e-cigarette use (AOR=1.32, 95% CI=1.17, 1.49). The risk of having had respiratory disease was also significantly associated with former combustible tobacco smoking (AOR=1.29, 95% CI=1.14, 1.47) and current combustible tobacco smoking (AOR=1.61, 95% CI=1.42, 1.82). Effects of e-cigarette and all combustible

tobacco use were independent risk factors for respiratory disease (variance inflation factors <1.2). Among people who did not report respiratory disease

at Wave 1, the longitudinal analysis revealed statistically

www.ajpmonline.org

Table 2. Respiratory Disease, Tobacco Use, Clinical and Demographic Variables

Variables Respiratory disease p-value

Wave 1 (n=32,320)

E-cigarette user Yes (n=5,457) No (n=26,646)

Never 3,123 (76.5) 17,511 (83.3) <0.001 Former 1,590 (16.1) 6,248 (11.5)

Current 744 (7.4) 2,887 (5.2)

Combustible tobacco smoker Yes (n=5,212) No (n=25,467)

Never 597 (22.0) 4,220 (29.7) <0.001 Former 1,684 (46.1) 8,689 (45.4)

Current 2,931 (31.9) 12,558 (24.9)

Cigarette smoker Yes (n=5,449) No (n=26,581)

Never 914 (25.9) 6,172 (34.5) <0.001 Former 1,848 (46.3) 9,689 (45.3)

Current 2,687 (27.9) 10,720 (20.2)

Wave 2 or 3a

E-cigarette user Yes (n=1,116) No (n=18,194)

Never 635 (74.1) 12,114 (83.7) <0.001 Former 314 (17.2) 4,188 (11.2)

Current 167 (8.7) 1,892 (5.1)

Combustible tobacco smoker Yes (n=1,069) No (n=17,464)

Never 110 (21.9) 2,995 (30.1) <0.001 Former 259 (36.8) 6,229 (46.1)

Current 700 (41.3) 8,240 (23.8)

Cigarette smoker Yes (n=1,114) No (n=18,152)

Never 170 (25.9) 4,313 (34.8) <0.001 Former 284 (37.0) 6,893 (46.1)

Current 660 (37.1) 6,946 (19.1)

Covariates at Wave 1

Demographic

Age in years <0.001 18−24 1,461 (13.3) 7,622 (12.9)

25−34 873 (14.4) 5,438 (18.3)

35−44 752 (14.0) 4,168 (17.0)

45−54 832 (16.2) 3,982 (18.2)

55−64 843 (18.5) 3,114 (16.3)

65−74 503 (14.8) 1,599 (10.4)

75 and above 202 (8.8) 781 (6.8)

BMI (§SD), kg/m2 29.4 (§8.1) 27.8 (§7.2) <0.001 Sex

Male 2,344 (40.9) 13,898 (49.4) <0.001 Female 3,122 (59.1) 12,811 (50.6)

Poverty level/income

Below poverty 1,954 (29.9) 7,950 (24.3) <0.001 At or above poverty 2,990 (70.1) 16,207 (75.7)

Race/ethnicity

White 3,991 (78.5) 19,795 (77.8) 0.326

Black 843 (12.6) 4,178 (12.3)

Other 632 (8.9) 2,736 (9.9)

Clinical status

High blood pressure

Yes 1,765 (39.1) 5,334 (25.8) <0.001 No 3,686 (60.9) 21,321 (74.2)

(continued on next page)

Bhatta and Glantz / Am J Prev Med 2020;58(2):182−190 185

February 2020

Table 2. Respiratory Disease, Tobacco Use, Clinical and Demographic Variables (continued)

Variables Respiratory disease p-value

High cholesterol

Yes 1,350 (31.2) 4,119 (21.5) <0.001 No 4,101 (68.8) 22,536 (78.5)

Diabetes mellitus

Yes 971 (21.9) 2,601 (12.6) <0.001 No 4,490 (78.1) 24,079 (87.4)

Note: Numbers in parentheses are weighted percentages or SDs. Chi-square analysis was used for counts and t-test for continuous variables. aExcluding respondents who had respiratory disease at Wave 1, n=19,475.

186 Bhatta and Glantz / Am J Prev Med 2020;58(2):182−190

significant associations between former e-cigarette use (AOR=1.31, 95% CI=1.07, 1.60) and current e-cigarette use (AOR=1.29, 95% CI=1.03, 1.61) at Wave 1 and hav- ing incident respiratory disease at Waves 2 or 3 adjusting for combustible tobacco smoking, demographic, and clinical variables. Current combustible tobacco smoking (AOR=2.56, 95% CI=1.92, 3.41) was also significantly associated with having respiratory disease at Waves 2 or 3 (Table 3, right columns). Effects of e-cigarette and all combustible tobacco use were independent risk factors for respiratory disease (all variance inflation factors <1.2). A supplemental analysis using cigarette smoking

instead of any combustible tobacco product smoking also yielded statistically significant associations between former e-cigarette use (AOR=1.24, 95% CI=1.03, 1.50) and current e-cigarette use (AOR=1.23, 95% CI=1.00, 1.51) at Wave 1 and having incident respiratory disease at Waves 2 or 3 adjusting for demographic and clinical variables (Appendix Table 5, available online). Among the former cigarette smokers, 79.2% quit >1 year ago, 17.1% reported quitting in the past year, and the remain- ing 3.2% reported quitting in the last 30 days. Current cigarette smoking (AOR=2.70, 95% CI=2.12, 3.45) was also significantly associated with having respiratory dis- ease at Waves 2 or 3. Effects of e-cigarette and conven- tional cigarette use were independent risk factors for respiratory disease (all variance inflation factors <1.2). Consistent with existing literature, this study found

increased risk of respiratory disease associated with hypertension27,28 and diabetes29 (Appendix Table 5, available online). E-cigarette use at Wave 1 was associated with elevated

point estimates of incidence of specific respiratory con- ditions (COPD, chronic bronchitis, emphysema, and asthma) at Waves 2 or 3. However, because of the small number of incidents at Wave 2 and 3, some of these point estimates did not reach statistical significance (Appendix Table 6, available online), which is why the

primary analysis combined all the respiratory conditions (i.e., to increase statistical power). Pooling conditions also avoids the problem of double counting, as some of these respiratory diseases tend to occur together. This study assessed the possibility of reverse causality

by estimating the odds of initiating e-cigarette use by Wave 2 or 3 combined as a function of having respira- tory disease at Wave 1 among people who had never used e-cigarettes at Wave 1 (Table 4). Having respiratory disease at Wave 1 significantly predicted future e-ciga- rette use (p<0.001).

DISCUSSION

This study is the first population-based longitudinal analysis of the association between e-cigarette use and incident respiratory disease, with current e-cigarette use elevating the odds of developing incident respiratory dis- ease by a factor of 1.29 (95% CI=1.03, 1.61) in the longi- tudinal sample. The risk of respiratory disease is independent of, and in addition to, the risks associated with current combustible tobacco smoking (AOR=2.56, 95% CI=1.92, 3.41), as well as cigarettes alone. This find- ing is consistent with what would be expected based on animal11−16 and human studies17−20 of the biological effects of e-cigarettes as well as cross-sectional studies of e-cigarette use and respiratory illness21−24 and a longitu- dinal study of people with COPD.25 The risks that were identified in this longitudinal analysis were similar to the risks found in the cross-sectional analysis of PATH Wave 1 for e-cigarettes (AOR=1.29 for current users in the longitudinal analysis vs AOR=1.32 in the cross-sec- tional analysis; Table 3). The point estimate of risk was lower than the AOR (1.86; 95% CI=1.22, 2.83) Perez et al. (E-cigarette use is associated with emphysema, chronic bronchitis and COPD. In: American Thoracic Society 2018 International Conference) reported for the cross-sectional risk of COPD (including chronic bron- chitis and emphysema), although the CIs overlap with

www.ajpmonline.org

Table 3. Associations Between E-Cigarette Use and Respiratory Disease

Cross-sectional associations between e-cigarette user and respiratory disease

at Wave 1 (baseline)

Longitudinal association between incident respiratory disease

(at Wave 2 or 3) and e-cigarette user at Wave 1 excluding people who reported respiratory disease at Wave 1

Variables AOR (95% CI) p-value AOR (95% CI) p-value

E-cigarette user

Never ref ref

Former 1.34 (1.23, 1.46) <0.001 1.31 (1.07, 1.60) 0.009

Current 1.32 (1.17, 1.49) <0.001 1.29 (1.03, 1.61) 0.026

Combustible tobacco smoker

Never ref ref

Former 1.29 (1.14, 1.47) <0.001 1.16 (0.87, 1.57) 0.315

Current 1.61 (1.42, 1.82) <0.001 2.56 (1.92, 3.41) <0.001 High blood pressure

Yes 1.40 (1.21, 1.61) <0.001 1.27 (1.02, 1.58) 0.033

High cholesterol

Yes 1.25 (1.11, 1.41) <0.001 1.04 (0.79, 1.38) 0.741

Diabetes mellitus

Yes 1.38 (1.20, 1.60) <0.001 1.30 (0.98, 1.72) 0.073

Age in years

18−24 ref ref

25−34 0.75 (0.67, 0.83) <0.001 0.65 (0.49, 0.87) 0.004

35−44 0.74 (0.65, 0.85) <0.001 1.05 (0.80, 1.38) 0.741

45−54 0.76 (0.66, 0.87) <0.001 1.37 (1.08, 1.74) 0.012

55−64 0.90 (0.76, 1.07) 0.242 1.33 (0.99, 1.78) 0.060

65−74 1.00 (0.84, 1.19) 0.993 1.22 (0.79, 1.88) 0.378

75 and above 1.05 (0.81, 1.36) 0.726 1.82 (1.02, 3.22) 0.044

BMI 1.02 (1.02, 1.03) <0.001 1.03 (1.02, 1.04) <0.001 Sex

Female 1.50 (1.37, 1.63) <0.001 1.72 (1.41, 2.09) <0.001 Poverty level

At or above poverty 0.80 (0.72, 0.89) <0.001 0.66 (0.54, 0.81) <0.001 Race/ethnicity

White ref ref

Black 0.89 (0.80, 1.01) 0.067 1.39 (1.13, 1.72) 0.003

Other 1.02 (0.85, 1.22) 0.837 1.15 (0.82, 2.11) 0.418

Sample size 32,320 19,475

VIF <1.2 <1.2 Note: Boldface indicates statistical significance (p<0.05). VIF, variance inflation factors.

Bhatta and Glantz / Am J Prev Med 2020;58(2):182−190 187

this study estimates. Rather than doing a multivariate analysis, Perez and colleagues used propensity score matching to control for smoking, secondhand smoke exposure, and other covariates. The finding that the effects of e-cigarettes and ciga-

rette smoking were independent risks is consistent with the evidence of substantial differences in the proteins expressed in human lung epithelial cells derived from smoker and e-cigarette user airways.20 Biomarker data from Wave 1 of PATH revealed higher levels of

February 2020

biomarkers of nicotine and toxicant exposure among dual users (e-cigarettes plus cigarettes) than smokers.30

Levels among e-cigarette−only users were higher than for people who smoked but below levels of cigarette smokers. Because the different products are independently

associated with risk of developing pulmonary disease, it is possible to use the results in Table 3 to estimate the risks of other behaviors, including dual use and switch- ing from combustible tobacco to e-cigarettes. For

Table 4. Reverse Causality Analysis: Longitudinal Predictors of Current E-Cigarette Use at Waves 2 or 3 as a Function of Reporting Respiratory Disease at Wave 1 Among Current Combustible Tobacco Smokers at Wave 1

Variables at Wave 1 AOR (95% CI) p-value

Respiratory disease

No ref

Yes 1.44 (1.22, 1.70) <0.001 High blood pressure

Yes 1.18 (0.95, 1.46) 0.130

High cholesterol

Yes 0.88 (0.74, 1.06) 0.174

Diabetes mellitus

Yes 1.16 (0.94, 1.44) 0.178

Age in years

18−24 ref

25−34 0.59 (0.47, 0.73) <0.001 35−44 0.43 (0.35, 0.53) <0.001 45−54 0.24 (0.19, 0.30) <0.001 55−64 0.18 (0.14, 0.23) <0.001 65−74 0.11 (0.07, 0.15) <0.001 75 and above 0.04 (0.01, 0.13) <0.001

BMI 0.99 (0.98, 1.00) 0.056

Sex

Female 1.46 (1.27, 1.68) <0.001 Poverty level/income

At or above poverty 0.92 (0.80, 1.05) 0.232

Race/ethnicity

White ref

Black 0.51 (0.42, 0.62) <0.001 Other 0.90 (0.68, 1.17) 0.427

VIF <1.2 Total sample size 11,192

Note: Boldface indicates statistical significance (p<0.05). Every day, some day, and current experimental users included. VIF, variance inflation factors.

188 Bhatta and Glantz / Am J Prev Med 2020;58(2):182−190

example, the total odds of developing respiratory disease among a former combustible tobacco smoker who cur- rently uses e-cigarettes is (odds of respiratory disease among former combustible tobacco smoker)£ (odds of respiratory disease among current e-cigarette user) = 1.16£ 1.29 = 1.50, compared with a never combustible tobacco smoker who has never used e-cigarettes. Thus, odds of developing respiratory disease for an individual who switched from combustible tobacco smoking to e- cigarette use would change by a factor of ([odds of respi- ratory disease among former combustible tobacco smoker]£ [odds of respiratory disease among current e-cigarette user])/(odds of respiratory disease among current combustible tobacco smoker) = (1.16£ 1.29)/ 2.56 = 0.58. This result suggests that switching from combustible tobacco to e-cigarettes would lower risk of developing respiratory disease, but among combustible

tobacco users who were not using e-cigarettes at Wave 1, only 0.9% of current e-cigarette users at Wave 2 and 0.8% at Wave 3 had switched exclusively to e-cigarettes. The numbers for cigarette smokers were 8.6% and 9.3%. The much more common pattern is dual use, in which

an e-cigarette user continues to smoke combusted tobacco products at the same time (93.7% of e-cigarette users at Wave 2 and 91.2% at Wave 3 also used combus- tible tobacco; 73.3% of e-cigarette users at Wave 2 and 64.9% at Wave 3 also smoked cigarettes). The total odds of developing respiratory disease for a current dual user is (odds of respiratory disease among current combusti- ble tobacco smoker)£ (odds of respiratory disease among current e-cigarette user) = 2.56£ 1.29 = 3.30 compared with a never smoker who never used e-ciga- rettes (which is similar to the direct estimate, AOR=3.04; Appendix Table 7, available online). The same situation applies to e-cigarettes and cigarettes (AOR=3.32). In other words, dual use of e-cigarettes and combustible tobacco (including cigarettes) is more dan- gerous than using either product alone. The major strength of this study is that it is based on a

large, nationally representative, randomly selected sam- ple of the population, with longitudinal follow-up. The longitudinal design allows much stronger conclusions about causality than in earlier cross-sectional studies (although this study found similar risks for e-cigarettes in longitudinal and cross-sectional analyses). Another strength of the longitudinal component of the study is that the incident cases of respiratory disease occurred many years after e-cigarettes entered the market and information on new diagnoses was collected within a year of respondents being informed of their diagnoses.

Limitations Several respiratory conditions were combined to obtain enough events to achieve adequate power. For the same reason, this study did not distinguish between daily and nondaily product use and included both established (smoked >100 cigarettes) and experimenters in the for- mer smoker group. There is a possibility of recall bias because use of e-

cigarettes, conventional cigarettes, and other combusti- ble tobacco products were self-reported as were clinical conditions. Participants with respiratory diseases might over-report e-cigarette, conventional cigarette, and other combustible tobacco use. There is also possibility of recall bias because doctor diagnoses of lung or respira- tory diseases is reported by respondents rather than being based on actual hospital records but the questions. However, the question Has a doctor or other health pro- fessional ever told you that you had any of the following lung or respiratory conditions: COPD, chronic bronchitis,

www.ajpmonline.org

Bhatta and Glantz / Am J Prev Med 2020;58(2):182−190 189

emphysema, and asthma? is used widely in epidemio- logic studies, including other federal surveys such as the National Health Interview Survey. This question has been validated against direct clinical observation in at least 2 studies; one reported that 98% of patients had clinically or spirometrically validated among self-reported diagnosis of COPD31 and another found clinical validation in 83%, 84%, and 90% of nurses self-reporting diagnoses of COPD.32 Research to validate analogous questions about myocardial infarction also found high agreement (81% −98%) with medical records.33,34 The longitudinal follow- up was only 2 years, but COPD has been detected in peo- ple after 1−9 years of smoking.35 In addition, this study examined incident cases, which may have been developing for some time before symptoms were manifest. The simi- larity of the cross-sectional and longitudinal estimates supports this idea. As noted above, this study found p<0.001 for reverse

causality, which could be consistent with a hypothesis that some individuals with respiratory disease try e-ciga- rettes believing they might be therapeutic. This study limited to control for intensity and type of e-cigarette use, which could affect the respiratory outcome. There is also always the possibility that other important con- founders were not measured in the PATH study.

CONCLUSIONS

Current use of e-cigarettes appears to be an independent risk factor for respiratory disease in addition to all com- bustible tobacco smoking. Although switching from combustible tobacco, including cigarettes, to e-cigarettes theoretically could reduce the risk of developing respira- tory disease, current evidence indicates a high prevalence of dual use, which is associated with increased risk beyond combustible tobacco use. In addition, for most smokers, using an e-cigarette is associated with lower odds of successfully quitting smoking.4,36 E-cigarettes should not be recommended.

ACKNOWLEDGMENTS This work was supported by grants R01DA043950 from the National Institute on Drug Abuse; P50CA180890 from the National Cancer Institute and the U.S. Food and Drug Adminis- tration Center for Tobacco Products; U54HL147127 from the National Heart, Lung, and Blood Institute and the Food and Drug Administration Center for Tobacco Products; and the Uni- versity of California, San Francisco Helen Diller Family Compre- hensive Cancer Center Global Cancer Program. The content is solely the responsibility of the authors and does not necessarily represent the official views of NIH or the Food and Drug Admin- istration. The funding agencies played no role in study design; collection, analysis, and interpretation of data; writing the report; or the decision to submit for publication.

February 2020

Author contributions: Concept and design, analysis or inter- pretation of data: DNB, SAG. Drafting of the manuscript: DNB. Critical revision of the manuscript: SAG.

No financial disclosures were reported by the authors of this paper.

SUPPLEMENTAL MATERIAL Supplemental materials associated with this article can be found in the online version at https://doi.org/10.1016/j. amepre.2019.07.028.

REFERENCES 1. Murphy SL, Xu J, Kochanek KD, Curtin SC, Arias E. Deaths: final data

for 2015. Natl Vital Stat Rep. 2017;66(6):1–75. www.cdc.gov/nchs/ data/nvsr/nvsr66/nvsr66_06.pdf. Accessed August 20, 2018.

Ta gg e dP 2. WHO. World health statistics 2008. www.who.int/gho/publications/ world_health_statistics/EN_WHS08_Full.pdf T a g g e d E n d. Published 2008. Accessed August 20, 2018.

3. GBD 2015 Chronic Respiratory Disease Collaborators. Global, regional, and national deaths, prevalence, disability-adjusted life years, and years lived with disability for chronic obstructive pulmonary dis- ease and asthma, 1990−2015: a systematic analysis for the Global Bur- den of Disease Study 2015. Lancet Respir Med. 2017;5(9):691–706. https://doi.org/10.1016/S2213-2600(17)30293-X.

4. Glantz SA, Bareham DW. E-cigarettes: use, effects on smoking, risks, and policy implications. Annu Rev Public Health. 2018;39(1):215–235. https://doi.org/10.1146/annurev-publhealth-040617-013757.

5. Wieslander G, Norb€ack D, Lindgren T. Experimental exposure to pro- pylene glycol mist in aviation emergency training: acute ocular and respiratory effects. Occup Environ Med. 2001;58(10):649–655. https:// doi.org/10.1136/oem.58.10.649.

6. Kreiss K, Gomaa A, Kullman G, Fedan K, Simoes EJ, Enright PL. Clin- ical bronchiolitis obliterans in workers at a microwave-popcorn plant. N Engl J Med. 2002;347(5):330–338. https://doi.org/10.1056/ NEJMoa020300.

7. Farsalinos KE, Kistler KA, Gillman G, Voudris V. Evaluation of elec- tronic cigarette liquids and aerosol for the presence of selected inhala- tion toxins. Nicotine Tob Res. 2015;17(2):168–174. https://doi.org/ 10.1093/ntr/ntu176.

8. Behar RZ, Luo W, Lin SC, et al. Distribution, quantification and toxic- ity of cinnamaldehyde in electronic cigarette refill fluids and aerosols. Tob Control. 2016;25(suppl 2):ii94–ii102. https://doi.org/10.1136/ tobaccocontrol-2016-053224.

9. Olmedo P, Goessler W, Tanda S, et al. Metal concentrations in e-ciga- rette liquid and aerosol samples: the contribution of metallic coils. Environ Health Perspect. 2018;126(2):027010. https://doi.org/10.1289/ EHP2175.

10. Williams M, Villarreal A, Bozhilov K, Lin S, Talbot P. Metal and sili- cate particles including nanoparticles are present in electronic ciga- rette cartomizer fluid and aerosol. PLOS ONE. 2013;8(3):e57987. https://doi.org/10.1371/journal.pone.0057987.

11. Chun LF, Moazed F, Calfee CS, Matthay MA, Gotts JE. Pulmonary toxicity of e-cigarettes. Am J Physiol Lung Cell Mol Physiol. 2017;313 (2):L193–L206. https://doi.org/10.1152/ajplung.00071.2017.

12. Moretto N, Volpi G, Pastore F, Facchinetti F. Acrolein effects in pul- monary cells: relevance to chronic obstructive pulmonary disease. Ann N Y Acad Sci. 2012;1259(1):39–46. https://doi.org/10.1111/ j.1749-6632.2012.06531.x.

13. Garcia-Arcos I, Geraghty P, Baumlin N, et al. Chronic electronic ciga- rette exposure in mice induces features of COPD in a nicotine-depen- dent manner. Thorax. 2016;71(12):1119–1129. https://doi.org/10.1136/ thoraxjnl-2015-208039.

190 Bhatta and Glantz / Am J Prev Med 2020;58(2):182−190

14. Reinikovaite V, Rodriguez IE, Karoor V, et al. The effects of electronic cigarette vapour on the lung: direct comparison to tobacco smoke. Eur Respir J. 2018;51(4):1701661. https://doi.org/10.1183/13993003.01661- 2017.

15. Sussan TE, Gajghate S, Thimmulappa RK, et al. Exposure to electronic cigarettes impairs pulmonary anti-bacterial and anti-viral defenses in a mouse model. PLOS ONE. 2015;10(2):e0116861. https://doi.org/ 10.1371/journal.pone.0116861.

16. Crotty Alexander LE, Drummond CA, Hepokoski M, et al. Chronic inhalation of e-cigarette vapor containing nicotine disrupts airway barrier function and induces systemic inflammation and multiorgan fibrosis in mice. Am J Physiol Regul Integr Comp Physiol. 2018;314(6): R834–R847. https://doi.org/10.1152/ajpregu.00270.2017.

17. Martin EM, Clapp PW, Rebuli ME, et al. E-cigarette use results in sup- pression of immune and inflammatory-response genes in nasal epithe- lial cells similar to cigarette smoke. Am J Physiol Lung Cell Mol Physiol. 2016;311(1):L135–L144. https://doi.org/10.1152/ajplung.00170.2016.

18. Miyashita L, Suri R, Grigg J. The effect of e-cigarettes on nasal platelet activating factor receptor (PAFR) expression. Am J Respir Crit Care Med. 2017;195:A1029.

19. Reidel B, Radicioni G, Clapp PW, et al. E-cigarette use causes a unique innate immune response in the lung, involving increased neutrophilic activation and altered mucin secretion. Am J Respir Crit Care Med. 2018;197(4):492–501. https://doi.org/10.1164/rccm.201708-1590OC.

20. Ghosh A, Coakley RC, Mascenik T, et al. Chronic e-cigarette exposure alters the human bronchial epithelial proteome. Am J Respir Crit Care Med. 2018;198(1):67–76. https://doi.org/10.1164/rccm.201710-2033OC.

21. Cho JH, Paik SY. Association between electronic cigarette use and asthma among high school students in South Korea. PLOS ONE. 2016;11(3):e0151022. https://doi.org/10.1371/journal.pone.0151022.

22. McConnell R, Barrington-Trimis JL, Wang K, et al. Electronic cigarette use and respiratory symptoms in adolescents. Am J Respir Crit Care Med. 2017;195(8):1043–1049. https://doi.org/10.1164/rccm.201604- 0804OC.

23. Wang MP, Ho SY, Leung LT, Lam TH. Electronic cigarette use and respi- ratory symptoms in Chinese adolescents in Hong Kong. JAMA Pediatr. 2016;170(1):89–91. https://doi.org/10.1001/jamapediatrics.2015.3024.

24. Wills TA, Pagano I, Williams RJ, Tam EK. E-cigarette use and respira- tory disorder in an adult sample. Drug Alcohol Depend. 2019;194:363– 370. https://doi.org/10.1016/j.drugalcdep.2018.10.004.

25. Bowler RP, Hansel NN, Jacobson S, et al. Electronic cigarette use in US adults at risk for or with COPD: analysis from two observational cohorts. J Gen Intern Med. 2017;32(12):1315–1322. https://doi.org/ 10.1007/s11606-017-4150-7.

26. Inter-university Consortium for Political and Social Research. Popula- tion Assessment of Tobacco and Health (PATH) Study [United States] Restricted-Use Files: User Guide. www.icpsr.umich.edu/icpsr- web/ICPSR/studies/36231/versions/V14/datasets/0/files/1253955/ downloadDoc/doc?path=/pcms/studies/0/3/6/2/36231/V14/files/ 1253955. Updated April 10, 2018. Accessed August 20, 2018.

27. Imaizumi Y, Eguchi K, Kario K. Lung disease and hypertension. Pulse (Basel). 2014;2(1−4):103–112. https://doi.org/10.1159/000381684.

28. Maclay JD, MacNee W. Cardiovascular disease in COPD: mecha- nisms. Chest. 2013;143(3):798–807. https://doi.org/10.1378/chest.12- 0938.

29. George C, Ducatman AM, Conway BN. Increased risk of respiratory diseases in adults with Type 1 and Type 2 diabetes. Diabetes Res Clin Pract. 2018;142:46–55. https://doi.org/10.1016/j.diabres.2018.05.029.

30. Goniewicz ML, Smith DM, Edwards KC, et al. Comparison of nicotine and toxicant exposure in users of electronic cigarettes and combustible cigarettes. JAMA Netw Open. 2018;1(8):e185937. https://doi.org/ 10.1001/jamanetworkopen.2018.5937.

31. Radeos MS, Cydulka RK, Rowe BH, Barr RG, Clark S, Camargo CA Jr. Validation of self-reported chronic obstructive pulmonary disease among patients in the ED. Am J Emerg Med. 2009;27(2):191–196. https://doi.org/10.1016/j.ajem.2008.01.011.

32. Barr RG, Herbstman J, Speizer FE, Camargo CA Jr. Validation of self- reported chronic obstructive pulmonary disease in a cohort study of nurses. Am J Epidemiol. 2002;155(10):965–971. https://doi.org/ 10.1093/aje/155.10.965.

33. Tretli S, Lund-Larsen PG, Foss OP. Reliability of questionnaire infor- mation on cardiovascular disease and diabetes: cardiovascular disease study in Finnmark County. J Epidemiol Commun Health. 1982;36 (4):269–273. https://doi.org/10.1136/jech.36.4.269.

34. Okura Y, Urban LH, Mahoney DW, Jacobsen SJ, Rodeheffer RJ. Agreement between self-report questionnaires and medical record data was substantial for diabetes, hypertension, myocardial infarc- tion and stroke but not for heart failure. J Clin Epidemiol. 2004;57(10):1096–1103. https://doi.org/10.1016/j.jclinepi.2004.04. 005.

35. Liu Y, Pleasants RA, Croft JB, et al. Smoking duration, respiratory symptoms, and COPD in adults aged >45 years with a smoking his- tory. Int J Chron Obstruct Pulmon Dis. 2015;10:1409–1416. https:// doi.org/10.2147/COPD.S82259.

36. Kalkhoran S, Glantz SA. E-cigarettes and smoking cessation in real- world and clinical settings: a systematic review and meta-analysis. Lancet Respir Med. 2016;4(2):116–128. https://doi.org/10.1016/S2213- 2600(15)00521-4.

www.ajpmonline.org