Reply to Commentary on “Adult Smokers’ Complete Switching Away from Cigarettes at 6, 9, and 12 Months After Initially Purchasing a JUUL e-Cigarette” and the Adult JUUL Switching and Smoking Trajectories (ADJUSST) Study

Abstract

Background

This is authors’ reply to the Commentary on our publication entitled “Adult smokers’ Complete Switching Away from Cigarettes at 6, 9, and 12 Months after Initially Purchasing a JUUL e-Cigarette.”

Methods and Results

Analyses addressed questions about follow-up rates and missed responses in the Adult JUUL Switching and Smoking Trajectories (ADJUSST) Study. Results demonstrate limited potential for selection bias, as participants who missed surveys were similar to those with complete data, and re-contact of participants who missed a follow-up indicated almost half were not smoking. Imputing smoking behavior for missing data would likely introduce bias and is not appropriate. The study demonstrated that JUUL products can facilitate high rates of complete switching away from cigarettes as suggested in previous experimental and observational studies. The ADJUSST cohort, including baseline nonsmokers, demonstrates a net reduction in smoking prevalence. Moreover, population modeling considering both benefits and harms demonstrated a net population benefit.

Conclusion

While the ADJUSST Study is not without limitations, the findings are consistent with multiple streams of real-world evidence that indicate that ENDS can facilitate switching among adults who smoke, and provide population benefits.

Keywords:

• JUUL
• Electronic nicotine delivery systems
• E-cigarettes
• Adult JUUL Switching and Smoking Trajectories (ADJUSST) Study

We appreciate the comments raised by Mantey et al., regarding our paper (Kim et al., 2024). We believe the comments fundamentally misinterpret this study and mischaracterize the scientific literature on electronic nicotine delivery systems (ENDS). We provide brief responses and clarifications.

Mantey et al.’s misunderstandings about the adult JUUL switching and smoking trajectories (ADJUSST) Study

Erroneously treating the ADJUSST study as a clinical trial for smoking cessation and assigning smoking behavior to missing observations

The commentary treats the ADJUSST study as though it were a clinical smoking-cessation treatment trial. However, ADJUSST is an observational study and very different from clinical trials such as Anthenelli et al. (Anthenelli et al., 2016) (Table 1). In treatment trials, the assumption that participants who missed an observation must have been smoking is justified by the evidence that most participants who were lost-to-follow-up were indeed smoking (Foulds et al., 1993). ADJUSST participants who missed a follow-up do not follow a similar pattern. When non-responders of the Month-12 survey were contacted again (Shiffman et al., 2021), 46% reported no smoking in the past 30 days. Moreover, there was no expectation or goal for participants to stop smoking, hence no motivation for participants to misreport or avoid reporting smoking. Accordingly, observational studies in general, including Population Assessment of Tobacco and Health (PATH) (Brouwer et al., 2022; Harlow et al., 2021; 2023; Kalkhoran et al., 2020) and National Health Interview Survey (Farsalinos & Niaura, 2019; Johnson et al., 2019), do not assign smoking status to missed follow-ups.

Table 1. Differences between clinical trials for smoking cessation and the ADJUSST study.

	Treatment studies for smoking cessation, including Anthenelli et al (Anthenelli et al., 2016)	The ADJUSST Study (Shiffman et al., 2021)
Participants	Individuals motivated to quit smoking	Individuals purchasing JUUL for the first time, for varied reasons
Plans to quit smoking	Typically required	Not a requirement for the study (majority were not planning to quit in the next 30 days (Goldenson et al., 2021))
Explicit behavioral objective	Smoking cessation	No study-wide behavioral goals; simply observing the changes in smoking and ENDS use over time
Product provision	Products provided gratis	No free JUUL products were provided; Participants purchased JUUL (or not) of their own accord
Counseling	Counseling provided	Naturalistic observational study without any intervention. No counseling or guidance on the use of JUUL (beyond product insert)
Follow-up patterns	Often lost for good, associated with relapse (Foulds et al., 1993)	Often “come and go” from one survey to the next; missed surveys attributed to survey logistics (Shiffman et al., 2021)

Uninformed interpretation of follow-up rates

It is a valid concern that if participants who missed follow-ups were particularly likely to be smoking, this would bias estimates of switching rates. But that is not the case, as demonstrated in a paper cited in the comment (Shiffman et al., 2021).

A follow-up survey of participants who missed a survey showed high switch rates

ADJUSST participants who missed the Month-12 survey were contacted and asked about their smoking status and reasons for missing the survey (Shiffman et al., 2021). Unlike in smoking cessation clinical trials, where all lost to follow-up are thought to have resumed smoking (perhaps having dropped out because of this) (Foulds et al., 1993), 46% of ADJUSST Month-12 non-responders reported not smoking—similar to those who completed the 12-month follow-up. Moreover, participants’ reasons for not completing the survey were rarely linked to smoking or JUUL use, but overwhelmingly to survey logistics (i.e., compensation being too low).

The lack of association between participant characteristics and missing follow-ups

If participants who missed follow-ups were persons particularly likely to continue smoking, they would demonstrate baseline characteristics associated with difficulty quitting. In fact, there were no substantial variations in sociodemographic factors and smoking history between those who completed all surveys, only some surveys, or no surveys (Kim et al., 2024; Shiffman et al., 2021). The small differences observed did not always imply those who missed surveys were more likely to smoke (e.g., lower cigarette dependence among those who missed follow-ups).

Initial outcomes not associated with missing follow-ups

As a further test of the potential for bias, we tested differences in the actual outcome (switching away from smoking) at the earliest time-point in relation to missed follow-ups later. There was no difference in Month-1 smoking between those who missed later surveys and those who missed none (Kim et al., 2024).

The evidence suggests the potential for bias due to differential follow-up is minimal. It was appropriate to treat missing data in ADJUSST the same way it is usually treated in other observational studies.

Distorted computations of recruitment response rates

Computing a ‘response’ rate to broad study recruitment materials is neither typical practice nor informative. Clinical studies are typically promoted by advertising via newspapers, fliers, and social media. The number of people who may have been exposed to the advertisement is typically not calculated in such mass communication media but may be in the hundreds of thousands, yielding extremely low putative “response rates.” Similarly for ADJUSST, it is impossible to determine how many invitations were seen, or seen by first-time purchasers (an eligibility criterion to participate in the study). Using this as a denominator for ‘response rates’ is not informative.

Misattributing nicotine as the source of smoking-related harms and mischaracterizing dual use behavior without taking accompanied changes in smoking into account

There is a scientific consensus that tobacco-related harm is overwhelmingly caused by exposure to harmful constituents in smoke from combusted products like cigarettes (National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health, 2014). Nicotine, while addictive and not risk-free, is not responsible for the detrimental effects of smoking (Gottlieb & Zeller, 2017). Indeed, health risks of dual use are proportional to the number of cigarettes smoked (Anic et al., 2022; Majeed et al., 2020; Smith et al., 2021; Wang et al., 2019).

The comment assumed dual use added ENDS to participants’ preexisting cigarette consumption. That is often not the case. Specifically, at 12 months, 60% of ADJUSST dual users had reduced their cigarette consumption by at least half, with mean reductions of over 80% (Selya et al., 2021). Evidence suggests that this is a meaningful reduction in exposure and likely risk (Cohen et al., 2021). In an experimental study, AWS who were assigned to use JUUL while reducing their cigarette consumption by at least 50% showed significant reductions in exposure to harmful chemicals, roughly half of those seen with tobacco abstinence. FDA has noted regarding such reductions that “ENDS may serve as a harm reduction approach if users… dramatically reduce combusted tobacco product use.” (emphasis added) (U.S. Food & Drug Administration, 2022).

Further, ADJUSST also suggests that dual use is a transitional state between smoking and switching, rather than a stationary behavior (Selya et al., 2021). One month after purchasing a first JUUL, 72% of AWS were dual users; 11 months later, 42% of those had switched completely away from smoking.

Evaluating the population health impact of JUUL products on the population as a whole

The ADJUSST Study assessed both the potential benefits conferred by switching among AWS, as well as the potential risks to adult users who formerly or never smoked. The balance between the two shapes the population health impact of ENDS. To address this, smoking was prospectively evaluated among all participants in the ADJUSST cohort (Prakash et al., 2021), including those who were not currently smoking, and a population modeling study was conducted to quantify the overall health impact (Wissmann et al., 2021).

Smoking behavior among adults who formerly or never smoked in the ADJUSST cohort

ADJUSST recruited first-time purchasers of JUUL, including some who were not smoking at baseline; however, approximately 90% of the sample had a history of smoking (Prakash et al., 2021), and most (74%) never-smoking adults were already using ENDS when they purchased JUUL and thus were not nicotine-naïve (Shiffman & Holt, 2021). Two published papers on adults who formerly and never smoked found that, while some reported smoking during the subsequent year, smoking was generally light and intermittent (Le et al., 2021; Shiffman & Holt, 2021). Additionally, more frequent use of JUUL was prospectively associated with a reduced likelihood of smoking (Le et al., 2021; Shiffman & Holt, 2021).

Importantly, even counting smoking among baseline nonsmokers, the percentage of the ADJUSST cohort who were smoking was reduced by more than half, from 65% (Baseline) to 31% (Month 12) (Prakash et al., 2021).

Modeling population impact

A population health model assessed the overall population impact of ENDS, including ENDS uptake by nonsmokers, and potential subsequent transitions to smoking. The model estimates that, over a century, the availability of ENDS like JUUL would avert 2.5 million premature deaths in the US (Wissmann et al., 2021). Net population benefits of ENDS have been consistently found in other modeling studies (Levy et al., 2021; Mendez & Warner, 2021; Wagner & Clifton, 2021).

Real-world evidence on complete switching to ENDS

The commentary asserts that the real-world effectiveness of ENDS for smoking discontinuation is unclear, particularly with respect to JUUL. JUUL has been shown to facilitate complete switching in a randomized trial conducted by academic researchers (Pulvers et al., 2020), with complete switching reported by 28% of AWS assigned to JUUL after 6 wk. Real-world users of JUUL, outside of experimental settings, have not only been the focus of the ADJUSST Study (Goldenson et al., 2021; Kim et al., 2024; Prakash et al., 2021) but also the JUUL Adult Smoker Switching Study (Shiffman & Hannon, 2023), both of which showed over 50% switching rates at Month 12. PATH data echo these findings, as AWS who use ENDS have significantly higher rates of smoking abstinence (Kasza et al., 2024) and lower rates of relapse (Klemperer et al., 2023) than those not using ENDS. At the population level, smoking prevalence (Levy et al., 2021) and cigarette sales (Selya et al., 2023) have declined at a rate even faster than what was projected before ENDS.

Conclusion

Like any study, ADJUSST has limitations, which have been extensively discussed in the current and prior publications (Kim et al., 2024; Shiffman et al., 2021). However, the current commentary misinterprets the study and does not take into account published analyses that address or refute the issues raised in the comment. We hope that our clarifications have offered a clearer perspective on the study and its findings.

Declaration of interests

SK and SS provide consulting services on tobacco harm reduction on an exclusive basis to Juul Labs, Inc. through PinneyAssociates, Inc. NIG is a full-time employee of Juul Labs, Inc. SS holds a patent for a novel nicotine medication that has not been developed or commercialized.

Additional information

Funding

This work was supported by Juul Labs, Inc.