Our calculations indicate that the cumulative incidence of EVALI was approximately three-fold greater among cannabis users in states with no legal access to cannabis versus cannabis users that reside in areas with legal adult-use cannabis marketplaces. This may have been due to legal access to regulated and/or tested products. These findings suggest that differences in state-level cannabis policies should be considered in future investigations into the causes of EVALI.

The EVALI outbreak has largely been attributed to the use of informally/illegally obtained cannabis oil vaping devices (CDC 2020; Blount et al. 2019b), and our findings add to existing evidence suggesting incident EVALI was higher in areas without legal access to cannabis (Wing et al. 2020). In areas with legal adult-use cannabis marketplaces, it is possible that more cannabis users obtained their vaping products through licensed retail channels, which may be less prone to selling potentially harmful products containing agents such as vitamin E acetate (Wing et al. 2020; Duffy et al. 2020; Taylor et al. 2019). While more research is needed on the health effects of constituents used in all vaping products (Werner et al. 2020; National Academies of Science Engineering and Medicine 2018), we speculate that legal environments may have offered potential for implementing product quality standards and other requirements determined by state regulatory bodies. In turn, this may have provided protection for consumers who purchased cannabis products through legal channels. Regulatory authorities may also impose fines or legal sanctions on producers and/or retailers found to be violating state-determined product protocols, which may act as a deterrent toward the use of additives such as vitamin E acetate in their products. While our analysis did not examine specific policy elements that may have mitigated the outbreak or exerted protective effects, we speculate that several may be at play. For instance, product standards and laboratory testing requirements for products authorized for retail sale may provide a layer of protection in preventing the addition of problematic constituents, and identifying problematic constituents or batches of product prior to the point-of-sale. Additionally, the existence of “seed to sale” tracking and inventory systems may be used to identify products sold through dispensaries that are subsequently found to be hazardous to users; such systems can be leveraged to report such information to the public. Examining policy elements such as safety provisions, along with others that impact consumer purchase behaviors (e.g., price and taxation of legal vaping products) (Hall et al. 2019; National Academies of Science Engineering and Medicine 2018; Smart and Pacula 2019), remains important areas for further study in the context of EVALI.

Alongside general gaps in knowledge about the health effects of using vaping products (National Academies of Science Engineering and Medicine 2018), it should be noted that the limitations placed on research into cannabis means that there are many unknowns concerning cannabis product safety, particularly for novel products (e.g., vaping products, edibles) (National Academies of Science Engineering and Medicine 2017). Additionally, the heterogeneity of policy elements and level of enforcement across states likely factor into what, if any, level of protection legal environments may actually afford (Hall et al. 2019; Smart and Pacula 2019). It is possible that policy elements that may protect against health issues such as EVALI may also have indirect or unintended adverse public health effects pertinent to other important considerations in the legalization of cannabis use. For instance, legalized policies have been shown to facilitate heavier consumption among heavy cannabis users, which may increase susceptibility to dependence and cannabis use disorder (Hall et al. 2019; National Academies of Science Engineering and Medicine 2017; Smart and Pacula 2019). Further, many legal environments have failed to suppress thriving black markets for cannabis (e.g., California), which consumers may continue to access in spite of legalization due to price, availability, convenience, or other factors (Hall et al. 2019; National Academies of Science Engineering and Medicine 2017; Smart and Pacula 2019). The effect of state actions in response to the outbreak is also worth noting. For example, during the outbreak, the state of Oregon obtained reports of legally obtained cannabis vaping products being a source of EVALI-related contaminants. In turn, the State issued a recommendation for cannabis retailers to review their products for potential safety concerns and post warning signs about potential dangers of using vaping devices (Flaccus 2019). Many states (e.g., Washington, Colorado) implemented bans on the use of vitamin E acetate in vaping products (Boudi et al. 2019), while others still (e.g., Massachusetts) issued complete bans on the use of vaping products for any substance during and following the outbreak (Lannan n.d.). The nature and interaction of specific policy elements implemented by each state, the strength and duration of policy elements, and the impact of black market product availability, consumer behavior, and government actions in response to the EVALI outbreak are all important areas for future research on this topic.

Although this piece represents a novel perspective on potential contributing factors into EVALI, there are limitations to consider. Primarily, we acknowledge that these estimates may be flawed due to the use of information available in the public domain. However, our searches exhibited reasonable concordance with categorical estimates provided by CDC, which lends credibility to the data used for analysis. Additionally, the use of NSDUH data from 2017 to 2018 as a reference population may result in a slight overestimation of cumulative incidence of EVALI, since prevalence of cannabis use has been increasing over time (Hall et al. 2019). Unfortunately, these were the most recent data available; future assessments should aim to replicate these figures with more recent data on the US cannabis user population at the state-level. Case reporting discrepancies between state health departments and the CDC, as well as misreporting of cannabis use by patients (Blount et al. 2019b), may also play a role. Further, there is an absence of national-level data to use to estimate the proportion of cannabis users that have used cannabis oil vaping products (including those obtained through legal or illegal channels). Cannabis can be consumed in many ways, is most often smoked, and is subject to extensive poly-use along with other substances, such as nicotine (National Academies of Science Engineering and Medicine 2017; Werner et al. 2020). Therefore, not all cannabis users to will engage with cannabis oil vaping products, which may further affect our calculations to an unknown degree. Our results may have also been influenced by enactment of state-level executive orders in the wake of the EVALI outbreak. For instance, Massachusetts enacted a ban on the sale of any vaping products shortly after the beginning of the peak of this outbreak (Lannan n.d.). A ban may have exerted a protective effect or drove users to obtain vaping products from the illicit market. While our calculations showed that Massachusetts exhibited similar overall rates of EVALI to places where cannabis was fully illegal, the myriad of contributing and intersecting factors that could potentially drive our findings (e.g., executive order issuance, state-level density of retail dispensaries) remain important areas for future research. Finally, our assumptions and analyses rely on the current evidence base related to EVALI, which is continually emerging. Future assessments should take these factors into account when identifying additional contributors to EVALI.