Demographics of Survey Participants
We received responses to surveys from 1198 individuals, of which 933 included complete data (Fig. 1). Mapping of survey response IP addresses demonstrated the international reach obtained by the social media advertisements utilized, with responses from 47 countries across 6 continents (Fig. 2). Reported ages of participants ranged from 13 to 70 years, with a mean of 26 years. Gender distribution was fairly even with 53% identifying as female, 46% as male, and 1% as non-binary or who refused to answer. The majority of respondents were Caucasian (62%), with the next highest percent being Asian (19%) and Asian Indian (6%; Table 1).
Table 1
Demographics of Survey Responders
| Nonsmoker | Conventional | E-cig User | Dual User | Overall | p value |
| n = 472 | n = 109 | n = 74 | n = 278 | n = 933 |
Age | | | | | | < 0.001 |
Mean (SD) | 22.6 (12.1) | 34.6 (11.8) | 18.4 (3.9) | 30.6 (11.9) | 26.1 (12.6) |
Gender | | | | | | |
Female | 265 (56.14%) | 70 (64.22%) | 49 (66.22%) | 112(40.29%) | 496 (53.16%) | < 0·001 |
Male | 200(42.37%) | 39 (35.78%) | 24 (32.43%) | 163 (58.63) | 426 (45.66%) |
Other | 7 (1.48%) | 0 (0%) | 1 (1.35%) | 3 (1.08%) | 11 (1.18%) |
Total | 472 (99.99%) | 109 (100%) | 74 (100%) | 278 (100%) | 933 (100%) |
Ethnicity | | | | | | |
Alaska Native | 1 (0.21%) | 2 (1.85%) | 0 (0%) | 0 (0%) | 3 (0.32%) | |
Asian Indian | 18 (3.84%) | 9 (8.33%) | 9 (12.16%) | 17 (6.14%) | 53 (5.71%) | |
Asian | 123 (26.23%) | 8 (7.41%) | 14 (18.92%) | 27 (9.75%) | 172 (18.53%) | < 0.001 |
African American | 7 (1.49%) | 3 (2.78%) | 0 (0%) | 5 (1.81%) | 15 (1.62%) | |
Pacific Islander | 4 (0.85%) | 0 (0%) | 1 (1.35%) | 1 (0.36%) | 6 (0.65%) |
White | 260 (55.44%) | 80 (74.07%) | 37 (50%) | 202 72.92%) | 579 (62.39%) |
Middle Eastern | 0 (0%) | 0 (0%) | 0 (0%) | 2 (0.72%) | 2 (0.22%) |
Hispanic | 7 (1.49%) | 0 (0%) | 2 (2.7%) | 4 (1.44%) | 13 (1.4%) |
Mixed | 38 (8.1%) | 6 (5.56%) | 10 (13.51%) | 15 (5.42%) | 69 (7.44%) |
Prefer Not To Answer | 11 (2.35%) | 0 (0%) | 1 (1.35%) | 4 (1.44%) | 16 (1.72%) |
Total | 469 (100%) | 108 (100%) | 74 (99.99%) | 277 (100%) | 928 (100%) |
*Comparisons among the 4 groups are all significant with p < 0.001. ANOVA test was used for age; and Chi-square test was used for gender and ethnicity. **Multivariable analysis on ethnicity and race was condensed to 3 groups – Asian, White (non-Hispanic), and Other (Mixed, Prefer not to Answer, American Indian, Pacific Islander, and Black (non-Hispanic). |
The overall cohort was divided into four inhalant groups: non-smokers/non-vapers (51%), conventional tobacco smokers (12%), e-cigarette users (8%), and dual users of e-cigarettes and conventional tobacco (30%; Table 1). This again highlights the fact that many adults in the general population use both e-cigarette and conventional tobacco. There were significant differences in age across inhalant groups (p < 0.001). Conventional tobacco users had the highest mean age (35 years, SD = 11.8), while e-cigarette users had the lowest mean age (18 years, SD = 3.9; Table 1). For dual users, the percentage of Caucasian respondents was highest at 73%, followed by Asian at 10%, and Asian Indian at 6%. E-cigarette users had a lower percentage of Caucasian respondents (50%) and higher mixed race at 14%. The racial ethnicity self-identification in the conventional tobacco user group was similar to dual users; however, non-smokers/ non-vapers had a lower percentage of Caucasian respondents (55%) and a higher percentage of Asian respondents (26%). There was a low response rate of Alaska Native, Pacific Islander, Middle Eastern, Hispanic, African American and mixed ethnicities.
Presence of Cough is Associated with Poorer Sleep Quality and Dual Inhalant Use
Analysis of PSQI scores by presence of cough revealed that subjects with a cough had higher PSQI scores (7.64, SD = 3.661), indicating worsened sleep quality, compared to those without a cough (6.88, SD = 3.648; p = 0.002). These data suggest that cough may have a direct adverse impact on sleep quality in this cohort. However, we didn’t find a significant association between the inhalant groups and PSQI scores (Table 2). There was also no evidence of interaction between inhalant groups and gender on PSQI scores (p = 0.89). When gender was assessed as a variable, males were found to have lower PSQI scores (better sleep quality) relative to females (mean difference of -0.87, 95%CI: -1.36 to -0.39, p < 0.001).
Table 2
Linear Regression Model to assess the association between inhalant groups and PSQI scores
Pairwise Group Comparisons | Mean Difference | 95% CI | Raw p-value | Adjusted p-value |
Dual vs Nonsmoker | 0.30 | (-0.28, 0.88) | 0.314 | 0.471 |
Dual vs E-cig | 0.07 | (-0.92, 1.06) | 0.888 | 0.888 |
Dual vs Conventional | 0.81 | (-0.19, 1.65) | 0.056 | 0.336 |
E-cig vs Nonsmoker | 0.23 | (-0.69, 1.41) | 0.627 | 0.752 |
E-cig vs Conventional | 0.74 | (-0.40, 1.41) | 0.203 | 0.428 |
Conventional vs Nonsmoker | -0.52 | (-1.33, 1.89) | 0.214 | 0.428 |
Model adjusted for age (mean difference = 0.001 per year increase, 95%CI: (-0.02, 0.02), p = 0.95), gender (Male vs Female mean difference=-0.87, 95%CI: (-1.36, -0.39), p < 0.001), Race (Asian vs White mean difference =-0.03 (-0.63, 0.58); Other vs White mean difference = 0.31 (-0.46, 1.09); p = 0.703), and Ethnicity (Hispanic vs Non-Hispanic mean difference =-0.690 (-1.51, 0.13); Unknown vs Non-Hispanic mean difference =-0.92 (-2.53, 0.69), p = 0.16). |
Inhalant use was associated with higher reporting of cough. Dual users had the highest reporting of cough in the last 30 days (48%), with e-cigarette users having the lowest reporting of cough of all inhalant users (24%). Logistic regression model adjusting for age and gender showed dual users had a higher presence of cough compared to non-smokers/non-vapers (OR = 1.50, 95% CI: 1.09 to 2.07, raw p = 0.012, adjusted p = 0.034) and e-cigarette users (OR = 3.00, 95% CI: 1.64 to 5.48, raw and adjusted p < 0.001). E-cigarette users reported a lower presence of cough than non-smokers/ non-vapers (OR = 0.50, 95% CI: 0.29 to 0.89, raw p = 0.017, adjusted p = 0.034) (Table 4). There is no interaction between gender and inhalant groups on reported incidence of cough (p = 0.85).
Dual Use and Female Gender are Associated with Longer Sleep Latency
Sleep latency was associated with inhalant type, with dual users having longer sleep latency compared to non-smoking/non-vaping subjects (mean difference of 4.7; 95% CI: 1.75 to 7.65, raw p = 0.002, adjusted p = 0.012) (Table 3). Gender was also independently associated with sleep latency, with males having shorter sleep latency compared to females (mean difference of -2.86, 95% CI: -5.35 to -0.37, p = 0.024), such that males in this cohort were found to fall asleep faster than females. There was no significant interaction effect between gender and inhalant groups on sleep latency (p = 0.41), with both male and female dual users reporting longer sleep latency (4.25 minutes longer relative to non-smoking/non-vaping subjects for males, and 4.82 minutes for females).
Table 3
Linear Regression Model to assess the association between inhalant groups and sleep latency
Pairwise Group Comparisons | Mean Difference | 95% CI | Raw p-value | Adjusted p-value |
Dual vs Nonsmoker | 4.10 | (1.14, 7.06) | 0.007 | 0.042 |
Dual vs E-cig | 4.73 | (-0.31, 9.78) | 0.066 | 0.194 |
Dual vs Conventional | 3.67 | (-0.67, 8.01) | 0.097 | 0.194 |
E-cig vs Nonsmoker | -0.63 | (-5.29, 4.02) | 0.789 | 0.843 |
E-cig vs Conventional | -1.06 | (-6.96, 4.84) | 0.724 | 0.843 |
Conventional vs Nonsmoker | 0.43 | (-3.82, 4.68) | 0.843 | 0.843 |
Model adjusted for age (mean difference = 0.04 per year increase, 95%CI: (-0.07, 0.15), p = 0.522) and gender (Male vs Female mean difference=-2.88, 95%CI: (-5.36, -0.39), p = 0.023), Race (Asian vs White mean difference =-3.83 (-6.93,-0.74); Other vs White mean difference=-0.13 (-4.08, 3.83), p = 0.05), and Ethnicity (Hispanic vs Non-Hispanic mean difference = -4.38 (-8.27, -0.12); Unknown vs Non-Hispanic mean difference=-4.71 (-13.25, 3.84), p = 0.09 ). |
Table 4
Logistic Regression Model to assess the association between inhalant groups and presence of cough in the past 30 days
Pairwise Group Comparisons | OR | 95% CI | Raw p-value | Adjusted p-value |
Dual vs Nonsmoker | 1.47 | (1.06, 2.04) | 0.019 | 0.038 |
Dual vs E-cig | 3.03 | (1.65, 5.56) | < 0.001 | < 0.001 |
Dual vs Conventional | 1.54 | (0.96, 2.46) | 0.073 | 0.088 |
E-cig vs Nonsmoker | 0.49 | (0.28, 0.86) | 0.013 | 0.038 |
E-cig vs Conventional | 0.51 | (0.25, 1.01) | 0.055 | 0.083 |
Conventional vs Nonsmoker | 0.96 | (0.61, 1.52) | 0.855 | 0.855 |
Model adjusted for age (OR = 0.99 per year increase, 95%CI: (0.97, 1.00), p = 0.019) and gender (Male vs Female OR = 1.29, 95%CI: (0.98, 1.70), p = 0.01), Race (Asian vs White mean difference = 0.83 (0.59, 1.17); Other vs White mean difference = 0.96 (0.62, 1.49), p = 0.55), and Ethnicity (Hispanic vs Non-Hispanic mean difference = 0.55 (0.34, 0.89); Unknown vs Non-Hispanic mean difference = 0.29 (0.09, 0.87), p = 0.07). |
Dual Users of E-cigarettes and Conventional Tobacco have Greater Drug Use
Cross-tabulation was used to compare drug use across inhalant use groups. Regardless of gender, more dual users (42%) and e-cigarette users (45%) reported marijuana use relative to non-smokers/non-vapers (5%) and conventional tobacco users (19%; p < 0.001) (Fig. 3A). Additionally, more dual users (10%) reported use of cocaine than non-smokers/non vapers (0.4%), conventional users (5%), and e-cigarette users (0%; p < 0.001) (Fig. 3B). E-cigarette users (4%) and dual users (4%) had higher methamphetamine use than conventional tobacco smokers (2%) and non-smokers/non-vapers (0.4%; p = 0.007). More dual users (4%) reported use of N,N-dimethyltryptamine (DMT) than non-smokers/non vapers (0.2%), conventional tobacco smokers (1%), and e-cigarette users (1%). Two participants reported heroin/morphine/fentanyl use, and both were dual users. Seven participants reported using prescription tablet opiates, 4 reported using recreational opiates, and 2 reported using PCP, with a majority of these being dual users.