For exclusive e-cigarette use frequency, we adopted previously employed criteria: infrequent use defined as 1–5 days; intermediate use defined as 6–29 days; and daily use defined as all 30 days . To assess concurrent use frequency, we assessed the distribution and quartiles of cannabis use, which yielded a six-group variable to classify students as: infrequent exclusive e-cigarette users ; intermediate exclusive e-cigarette users ; daily exclusive ecigarette users ; infrequent concurrent e-cigarette and cannabis users ; intermediate concurrent users ; and frequent concurrent users . We calculated descriptive statistics for all variables of interest. To assess potential differences in current e-cigarette and cannabis use based on covariates, we performed a series of chi-square tests for the categorical covariates  and an independent t test for the continuous covariate . Then, a series of multivariable logistic regression models were fitted to explore the associations between exclusive ecigarette use and concurrent e-cigarette and cannabis use and COVID-19 symptoms, testing, and diagnosis. All models controlled for demographics, university site, fraternity/sorority membership, residence, combustible cigarette smoking, cigar smoking, and smokeless tobacco use. To assess frequency of e-cigarette and cannabis use, we built three similar logistic regression models adjusting for the covariates. We present adjusted odds ratios  and 95%CIs. Missing data were removed prior to analyses, and all analyses were two-tailed with statistical significance set at p < 0.05 and performed using Stata SE version 16 . This study provides evidence college student e-cigarette users who concurrently use cannabis in the past 30-days are at greater likelihood of experiencing COVID-19 symptoms and having a positive COVID-19 diagnosis, compared with exclusive e-cigarette users.

Confirming our hypothesis, frequency of concurrent e-cigarette and cannabis use was associated with increased odds of COVID-19 symptoms and diagnosis, with more pronounced odds observed as frequency of use groups increased, independent of student demographics and current use of combustible cigarettes, cigars, and smokeless tobacco. Thus, there appears to be a dose-related relationship, such that as use increased so too did the risk of experiencing COVID-19 symptoms and receiving a positive diagnosis. Specifically, for COVID-19 symptoms, effect size estimates were 3.5-fold among concurrent e-cigarette and cannabis grow equipment users at any frequency of use, and these estimates ranged from nearly 5-fold to 7.5-fold among infrequent, intermediate, and frequent concurrent users. Similar findings were indicated for COVID-19 diagnosis, with odds of nearly two times for concurrent users at any frequency of use, and approximately a 3-fold increase among both intermediate and frequent concurrent users. There are several potential explanations of why concurrent e-cigarette and cannabis users, especially those with more frequent use patterns, were at higher risk of experiencing COVID-19 symptoms when compared with exclusive e-cigarette users. First, combustible cannabis and tobacco smoke contain similar carcinogenic and other harmful chemical toxins, but cannabis topography results in higher tar and gas per-puff exposures than that of combustible tobacco smoke . This can lead to acute respiratory health symptoms , and potentially airway inflammation and infection especially among heavy or long-term cannabis users . Second, e-liquids of nicotine- and THC-containing vaping products vary in constituents and are a potential source of inhaled toxic metal exposure , and there are over 400 brands that provide diverse products . THC-containing e-liquids may be distinct from nicotine-containing e-liquids and can lead to higher respiratory illness likely due to varying inhaled chemical constituents . For example, it is important to note e-cigarette, or vaping, product use-associated lung injury  was linked to illicit THC-containing vaping products and vitamin E acetate in nearly all  of cases, with median EVALI case patient age of 23 years and the majority being male For these and other reasons, the Centers for Disease Control and Prevention recommends individuals not use THC-containing vaping products due to the potential of tampering with e-liquids .

While law enforcement seized vaping products containing vitamin E acetate intended for the illicit market , the clinical manifestations and symptoms of EVALI and COVID-19 and other respiratory illness overlap . Further research is needed to assess the associations of e-cigarette and cannabis use with COVID-19 outcomes based on use patterns including cannabis inhalation route, and device type and ingredients among vapers. Current smokeless tobacco use increased student e-cigarette users’ odds by nearly 3-fold for reporting COVID-19 symptoms, which aligns with previous research documenting increased risk of respiratory symptoms from smokeless tobacco use . Combustible cigarette smoking and cigar smoking were not significant covariates of COVID-19 symptoms, despite prior research linking dual ecigarette and combustible cigarette use with increased self-reported respiratory symptoms compared to exclusive e-cigarette use . Additionally, no differences were found based on current combustible cigarette, cigar, or smokeless tobacco use and COVID-19 diagnosis. Prior research indicates all forms of tobacco use may increase COVID-19 infection susceptibility via the ACE2 receptor  and the furin enzyme found in oral mucosa , and has been recognized as a risk factor for severe COVID-19 manifestations . Future research using objective measures is warranted to better understand the complex associations between tobacco product type and COVID-19-related outcomes. As posited, no differences were detected between current use groups and COVID-19 testing, likely based on similar random testing policies at each university during the data collection period. Concerning our findings on COVID-19 diagnosis, the active ingredients of THC and nicotine and toxic substances vary among cannabis and e-cigarette products, respectively, and cannabis chemicals are metabolized slower in the body, placing cannabis users at increased risk of COVID-19 infection . Interestingly, this study found those who were male and White had the highest percentages of having a COVID-19 diagnosis. Notably, the literature indicates the highest prevalence of EVALI cases are among those who are male and White .

Although this diagnosis was not assessed in this study, future work should examine the associations of e-cigarette and cannabis use and COVID-19 diagnoses and other specific diagnoses such as EVALI or pneumonia. Other explanations for higher odds of concurrent users having a COVID-19 diagnosis are behavior-related, including tendencies of sharing devices with others and hand-to-lip contact while using these products , 2021, which also increases COVID-19 risk via contact and fomite transmission . About 1-in-2 young adult lifetime e-cigarette users report sharing devices with others , which may explain this study’s finding that fraternity/sorority members had a higher likelihood of reporting a COVID-19 diagnosis. Moreover, e-cigarette use may ultimately increase risk-taking behaviors during young adulthood, including but not limited to concurrent cannabis use . Research indicates dual e-cigarette and combustible cigarette use is associated with poor compliance with COVID-19- related social distancing behaviors . Thus, it is highly likely e-cigarette users who engage in concurrent use of cannabis did not engage in recommended preventive health behaviors . While this study has several strengths, limitations should be noted. First, while students were enrolled at four geographically diverse universities across the U.S., our cross-sectional sample is not nationally representative and therefore our results are not generalizable to all U.S. student e-cigarette users. Longitudinal research is needed to assess causal associations between e-cigarette and cannabis use patterns with COVID-19-related outcomes. In a similar vein, we were unable to objectively measure COVID-19 symptoms, testing, and diagnosis. For example, the survey language specifically asked whether students were currently experiencing any COVID-19 symptoms from the Centers for Disease Control and Prevention’s COVID-19 symptoms list ; but since some of the symptoms were nonspecific to COVID-19, students may have reported a symptom  while not having COVID-19 concerns. Additionally, since our student sample included those who currently used e-cigarettes, we were unable to compare exclusive e-cigarette use versus non-use nor exclusive cannabis use. We assessed self-reported e-cigarette use frequency in number of days and cannabis use frequency in number of times used in the past 30 days and used categorical cut points to minimize the potential for recall bias.

We used standard national survey question language  to collect data on past 30-day cannabis use frequency based on number of times . Thus, we did not collect cannabis use frequency in number of days, and suggest this as a measure to be used in further research. Additionally, we did not collect information on cannabis use route . Future research should consider the use of biomarkers , and THC carboxylic acid the major metabolite of delta-9-tetrahydrocannabinol and patient medical records to cross-validate self-reported responses. Additionally, studies should take into consideration overall preventive health behaviors  and statewide and local policies  that may reduce infectious disease risk. Due to our recruitment methods , we could not calculate response rates, which resulted in varying participation rates that may have biased the sample. All four university campuses remained “open” during data collection. While we did not collect course engagement data ,vertical grow system future research should account for frequency of in-person class participation, which may have increased COVID-19 exposure and susceptibility. We did not have access to information on COVID-19 random testing rates at each university. COVID-19 testing rates may have varied at each campus based on test availability and accessibility on and off campuses . Cannabis is one of the most commonly used drugs worldwide  and young adults report some of the highest past-year rates of cannabis use. For example, data from South America indicates that around 14% of young adults in Argentina and 18% in Uruguay reported past-year cannabis use . In Europe, 19.1% of young adults in Spain and 13.4% in the United Kingdom  consumed cannabis during the last year.Studies from Canada show past 12-month cannabis use prevalence of 44% in young adults aged 16–19, 52% aged 20–24, and 24% aged 25 years or older . While in the U.S., the prevalence was 27% in young adults aged 18–34.Among young adults, college students are a specific high-risk subgroup. For example, annual prevalence of cannabis use is at historic high among U.S. college students  and daily cannabis consumption increased among U.S. university students in 2019 to 5.9% . Furthermore, college students who engage in a high-intensity or high-frequency pattern of use, are at greater risk of experiencing negative consequences , including addiction . Therefore, it is necessary to develop assessment measures to screen for problematic cannabis consumption among college students in order to increase identification and treatment of at risk students .

In a relevant review, among all potential instruments that assess cannabis-related problems, the Cannabis Use Disorders Identification Test  was selected as one of the most appropriate instruments for use in general population surveys because is simple and easy to understand, is brief and available in a public domain, encompasses a broad spectrum of cannabis-related problems and has been validated in general population samples and in samples of adolescents and young adults . The CUDIT was developed by Adamson and Sellman  based on the Alcohol Use Disorders Identification Test  in a cannabis-using alcohol-dependent sample . The questionnaire was later revised and improved  using a higher sample size of clinical patients . The most updated version of the CUDIT-R suggests a one-factor solution composed of 8 items, assessing: consumption , cannabis problems , physical dependence , and psychological features . Scores can range from 0 to 32, with a cut-off score of 13 indicative of a probable DSM-IV diagnosis of CUD Compared with the CUDIT , the CUDIT-R  has shown equivalent internal consistency , improved discriminant validity ; and an improved test–retest reliability index . To our knowledge, only two recent studies have provided validity and reliability evidence of the CUDIT-R scores among college students. Schultz et al. , in a sample of 229 undergraduates from the U.S. who reported past 30-day cannabis consumption, found good internal consistency of the questionnaire  and concurrent validity with cannabis related outcomes. They also found that a cut-off of six was adequate to differentiate between college students with and without problematic cannabis use. Risi, Sokolovsky, White, and Jackson , in a sample of 1,390 undergraduates from the U.S., found a one-factor structure for the CUDIT-R and configural and metric invariance across gender. Despite high rates of cannabis use globally, minimal research has examined the evidence of validity and reliability of the CUDIT-R among college students outside the U.S.