Sample
Weighted sample characteristics are shown in Table 1 (unweighted in Additional file 1: Table S1). By design, it comprised similar proportions from all three waves, age groups, sex and country. About two thirds considered themselves white, and about one quarter reported their family not or just being able to meet basic expenses. In the past 30 days, most had not smoked or vaped, few had smoked or vaped on at least 20 days, and 16.3% had used one or more other inhaled product, mostly cannabis (12.2%; 6.6% other tobacco; 0.9% heated tobacco).
Overall, 27.8% (n = 10,980) reported any respiratory symptoms in the past week, most commonly cough (16.1%), followed by shortness of breath (10.0%), chest pain (10.0%), phlegm (8.5%) and wheezing (2.6%). Among those who had not used any other products in the past month, 24.4% (n = 8049) reported any symptoms. There was some variation by socio-demographics, smoking, vaping and other product use (Table 1, Additional file 1: Table S2). In the 2021 waves, 3.3% thought they had had COVID-19 in the past 2 weeks, 2.8% did not know and 0.3% refused to say. Among those reporting COVID-19, 72.3% reported respiratory symptoms.
Among those who had vaped in the past 30 days (unweighted n = 4644), the median number of days vaped was 12 (interquartile range 3–30). Among those who had vaped and were aware of nicotine salts (unweighted n = 1874), just over half (53.5%) reported using salts; among all those who had vaped in the past 30 days, this was 21.6%. Cartridge/pod device types and fruit flavours were the most used, multiple devices and tobacco flavours the least (Table 2, unweighted in Additional file 1: Table S3).
Hypothesis 1. Past-30-day smoking and/or vaping
Compared with those who had only vaped, those who had both smoked and vaped had higher odds of experiencing any symptoms, whereas those who had neither smoked nor vaped had lower odds of symptoms and those who had only smoked were similar to those who had only vaped. These associations were found in unadjusted and adjusted analyses (Table 3).
Sensitivity analyses including only those who had not used other inhaled products found no difference between those who had both smoked and vaped compared with those who had only vaped; the other associations remained similar (Additional file 1: Table S4). Results from sensitivity analyses in the latter two waves and adjusting for COVID-19 remained very similar to the primary analysis (Additional file 1: Table S4), and associations with individual symptoms generally also agreed (Additional file 1: Table S5).
Hypothesis 2. Lifetime/current vaping
Compared with those who had never vaped, all other groups were more likely to have experienced respiratory symptoms, with those who had vaped on at least 20 of the past 30 days most likely to report symptoms, in unadjusted and adjusted analysis (Table 3). Results from sensitivity analyses including only those who had not used other inhaled products, or adjusting for COVID-19, were also similar (Additional file 1: Table S4), as were most associations with individual symptoms, except that wheezing differed less between groups (Additional file 1: Table S5).
Among those who had vaped in the past 30 days, each additional day vaping was associated with a small but statistically significant increase in the odds of experiencing respiratory symptoms (unadjusted OR (95% CI): 1.02 (1.01–1.02), p p
Hypothesis 3. Nicotine type
Among respondents who had vaped in the past 30 days, usual nicotine salt use was associated with higher odds of respiratory symptoms, and this association remained when adjusting for other variables (Table 4) and in the sensitivity analyses excluding those using other products or adjusting for COVID-19 (Additional file 1: Table S6). When including only those reporting awareness of nicotine salts, an association with symptoms was found in the unadjusted analyses [OR (95% CI): 1.21 (1.00–1.46), p = 0.047], but not in adjusted analysis [OR (95% CI): 1.14 (0.93–1.38), p = 0.205, unweighted n = 1874]. Those categorised as using nicotine salts had higher odds of reporting each of the five symptoms (Additional file 1: Table S7).
Hypothesis 4. Flavours and devices
Compared with tobacco flavours, menthol/mint and sweet flavours were associated with similar odds of symptoms; use of fruit flavours, ‘other/unknown’ and multiple flavours were all associated with higher odds of symptoms, and adjustment for other variables had little effect (Table 4). In the sensitivity analysis excluding those who had used other products, only use of multiple flavour types was associated with higher odds of symptoms (Additional file 1: Table S6). In the sensitivity analysis including COVID-19, in addition to fruit, other/unknown and multiple flavours, sweets flavours were also associated with higher odds of reporting symptoms than tobacco flavours (Additional file 1: Table S6). In analysis of individual symptoms, compared with tobacco flavours, all flavour types were associated with higher odds of cough, and ‘other/unknown’ flavour types were also associated with higher odds for chest pain; there were no significant associations with shortness of breath, wheezing or phlegm (Additional file 1: Table S7).
Compared with pre-filled cartridge/pod models, use of disposable, multiple or unknown device types was associated with higher odds of symptoms only in unadjusted analysis, and the odds for tank device types were similar to those for pre-filled cartridge/pod devices (Table 4). The sensitivity analyses found no significant associations (Additional file 1: Table S6). Compared with pre-filled cartridge/pod devices, disposable devices were associated with higher odds of shortness of breath, chest pain and phlegm; using multiple/unknown devices was also associated with higher odds of chest pain; for tanks, odds of each symptom were not different from pre-filled cartridge/pods. For wheezing and cough, there were no significant associations with device type (Additional file 1: Table S7).
Hypothesis 5. Country
The odds of symptoms were lower for Canada and higher for the US than for England among all respondents; among youth who had vaped in the past 30 days, the three countries were similar (Table 5). Sensitivity analyses were similar (Additional file 1: Table S8).
We found an interaction for country and past-30-day smoking and/or vaping (overall p
There was also a significant interaction for lifetime/current vaping (p
There was an overall significant interaction for nicotine type and country (p = 0.003), but none of the individual contrasts indicated a difference. The overall interactions for flavours (p = 0.257) or devices (p = 0.226) were not significant.
For individual symptoms, compared with England, Canada was associated with lower odds for all five symptoms, the US with lower odds for wheezing and higher odds for chest pain and cough. In those who had vaped in the past 30 days, the only association was for wheezing, with lower odds in the US (Additional file 1: Table S10).