In this study of more than 10,000 outpatients seeking triage for respiratory symptoms, active nicotine, cannabis, or multiple product use was independently associated with a higher likelihood of reporting dyspnea, even after accounting for chronic disease status. Overall, active users of nicotine or cannabis products reported a similar constellation of mild and moderate symptoms as non-users, but at higher rates. Active users were also more likely to report having comorbid chronic conditions, which may partially explain increased symptom rates. Prior studies have shown the presence of dyspnea to be a predictor of severe COVID-19 infection and disease progression (Shi et al., 2020; Zheng et al., 2020). These findings are consistent with existing literature, which demonstrates an association between cigarette smoking and increased upper respiratory infection symptomatology (Benowitz et al., 2022; Zhou et al., 2018). This may in part be due to the baseline increased frequency of symptoms, particularly respiratory symptoms, among those who smoke and vaporize nicotine products compared to their non-user counterparts (Sargent et al., 2022; Varella et al., 2022). The increased symptomatology may also be partially explained by the increased rates of chronic disease among users.

Respiratory illness symptom checkers generally do not account for smoking status (Coronavirus (COVID-19) Symptom Checker - Check Your Symptoms, 2023; COVID-19 Symptom Checker | VCU Health, 2023; Symptom Checker, 2023; Symptom Checker - St. Luke's's, 2023). However, our research suggests that 1) mild and moderate symptoms were more commonly reported in people actively using nicotine and/or cannabis compared to non-users, and 2) people actively vaping or dually smoking and vaping tobacco or cannabis products reported more dyspnea than non-users even after adjusting for age, race, comorbidities and other factors. These findings suggest that the use of vaporizers for tobacco or cannabis may not be associated with a difference in self-reported respiratory symptoms compared to smoking tobacco. These findings suggest that smoking and vaping status should be included with other comorbidities in respiratory illness symptom checkers and triage tools, both because it provides context for the symptoms reported, and because it may be a predictor of disease severity. COVID-19 is one of many global respiratory virus outbreaks over the past few decades, and the emergence of novel respiratory virus pandemics in the future is inevitable if historic trends continue (Noor and Maniha, 2020). If smoking and vaping measures are added to symptom checkers, future studies could further address the interplay between smoking and vaporizer use status and symptom reports, and further address associations between product use and clinical outcomes.

Importantly, we found increased respiratory symptoms not just among those smoking cigarettes, but also those using e-cigarettes, smoking cannabis and using cannabis vaporizers. While patients often perceive they may be mitigating the risk of adverse health outcomes by using vaporizers instead of smoking, the patterns of symptoms observed in this study, including severe symptoms, were similar between smoking and vaporizer users. There have been some studies suggesting a protective effect of active smoking for COVID-19 infection (Prinelli et al., 2021). However, this association may be spurious if people who smoke have more symptoms leading to higher rates of COVID-19 testing, and thus a larger proportion of negative tests. The high rates of symptoms among people who smoke and vape in this study are consistent with that explanation. The comorbidities associated with the reporting of potentially more severe symptoms in this study are consistent with previously published literature (Grasselli et al., 2020; Mahamat-Saleh et al., 2021; Richardson et al., 2020), including finding that people who smoke tend to have potentially more severe upper respiratory infections and worse outcomes (Benowitz et al., 2022; Jiang et al., 2020). This study also found that symptoms were common among users of all forms of nicotine and cannabis. Users may perceive that they are decreasing potential health risks by using cannabis instead of nicotine, or by vaping instead of smoking (Bhatia et al., 2022). However, this study shows a similar symptom distribution among all tobacco and cannabis users, and an increased association with dyspnea, which may be a marker of more potentially severe respiratory illness.

This study has several limitations. The COVID-19 & Flu Symptom Checker does not ask respondents about the frequency of use of nicotine and cannabis smoking and vaporized products, so we could not establish whether a dose-dependent association between smoking status and symptoms existed. In addition, the Symptom Checker does not assess for smoking history (i.e. pack-years of smoking), or recent smoking, and only was worded in present tense for “current smoking” without a specific time frame which could lead to misclassification. Since people who are vaping have possibly switched from cigarettes, there may be residual confounding from prior smoking history. Future studies may address this limitation with inclusion of a detailed smoking and vaping history built in as part of the triage tool for better sensitivity. The cross-sectional nature of this study also may not account for an individual's full symptom profile or disease severity, as respondents may have developed further symptoms after use of this triage tool. Both modes of use and type of products are potentially associated with respiratory outcomes, which were separated in our analyses (Bhat et al., 2023). Additionally, we used the reported symptom of dyspnea as a proxy for severity, rather than outcomes like hospitalization or death, because those outcomes were not available in our data set. Because this study was conducted somewhat early in the pandemic, most respondents were unvaccinated. Therefore, it is unknown whether the interaction between smoking status and respiratory viral symptoms would be different in the current environment. Our analytic sample was taken from a patient population which excluded those with who had unique clinical circumstances or reported emergent symptoms requiring immediate medical evaluation, with a prevalence of tobacco use about 7%, lower than the national estimates of ∼20%, but fairly consistent with California rates of cigarette smoking (6.3%) and vaping (3.5%) in 2020–2021 (California Department of Public Health, 2023). Differences due to self-selection of patients into the UCSF symptom checker tool or our patient population may result in decreased generalizability to other parts of the United States. Additionally, this study only includes adult participants, so may not be generalizable to pediatric populations. Lastly, given that all measures were self-reported in the tool, our results may be affected by measurement or recall bias, such as undiagnosed comorbidities. While many studies rely on self-reported measures of tobacco use with high reliability, validating self-reports on symptom checkers with EHR records may increase methodological rigor, reduce bias, and improve reliability in real-world practice (Morales et al., 2013).