Varenicline Over-The-Counter Trial on Efficacy and Safety

Abstract

Introduction

Varenicline helps people who smoke quit at rates 2–3 times greater than placebo. Currently, in the United States, varenicline is not available over the counter (OTC).

Aims and Methods

In this study, we assessed the safety and efficacy of 1 mg and 0.5 mg varenicline as an OTC medication for smoking cessation in comparison to placebo. This randomized, double-blind, placebo-controlled study was performed at two clinical sites in the United States of n = 313 people. The treatment period was 12 weeks. During the COVID pandemic, the protocol was modified to allow remote participation; verification of smoking status was via breath carbon monoxide levels for in-person visits and mailed urine cotinine kits for the remote participants.

Results

There was no difference in biologically confirmed continuous abstinence by condition between weeks 8 and 12; however, the odds of biologically confirmed point prevalence abstinence were higher for those in the 1 mg b.i.d. condition than for those in the placebo condition at week 12 (OR 3.39; 95% CI 1.49, 7.71), and were higher for those assigned to the 1.0 mg b.i.d. condition than the 0.5 mg b.i.d. condition at week 12 (OR 2.37; 95% CI 1.11, 5.05). Adverse events were modest, and as expected (vivid dreams and nausea in the medication conditions).

Conclusions

The results are suggestive that varenicline is safe and effective as an OTC medication.

Introduction

Despite decades of decreasing prevalence, cigarette smoking remains the leading cause of preventable death in the United States.¹ In 2021, 11.5% of US adults currently smoked cigarettes,² down nearly 40% in the past 20 years³; however, cigarette smoking prevalence remains high among those with the least education, lowest incomes, and least access to healthcare, among other individual characteristics.^1,2,4,5 Increasing access to and use of smoking cessation medications is a public health priority given this enduring disparity.⁶

In combination with behavioral support, the most effective smoking cessation medication available in the United States is varenicline^7,8; however, the most recent US data available suggest that only 15.5% of people who smoke used any prescription medication during their last quit attempt.⁹ Low utilization may be in part due to the inclusion of a US Food and Drug Administration-mandated “boxed” warning on varenicline packaging in 2009, driven by concerns that the medication increased risk of suicidal ideation and cardiovascular events.¹⁰ FDA removed the warning in 2016 after results from an FDA-mandated study (EAGLES) found no relationship between varenicline use and neuropsychiatric or cardiovascular adverse events.^11–13 Additionally, EAGLES results suggested that varenicline’s safety profile was similar to that of over-the-counter (OTC) nicotine replacement therapies but was significantly more effective at promoting cessation.^11,12

Given varenicline’s safety and efficacy profile, as well as the continued public health imperative to expand access to smoking cessation treatment, an examination of its potential as an OTC medication is warranted. Only two studies have assessed varenicline efficacy when provided without a prescription. In one of them, participants who smoked were assessed over the phone and then sent varenicline if eligible along with phone counseling.¹⁴ The other was a small in-person study that provided people who smoke with starter packs of varenicline.¹⁵ Both studies resulted in outcomes suggesting the medication’s promise in an OTC environment. Neither study employed a placebo condition or included biochemical verification of smoking status, which are criteria required to receive approval by the FDA.¹⁶ As of Spring 2024, varenicline is only accessible by prescription from a health care provider in the United States.

As varenicline is now, nicotine replacement therapies such as the nicotine patch and gum were also originally approved as prescription products. NRTs were approved for OTC sale after studies demonstrated their safety and efficacy without a prescription or guidance from a health care professional.^16,17 Given varenicline’s safety and efficacy in a prescription-only environment, it is warranted to assess varenicline’s safety and efficacy in an environment where the barrier to prescription-only use is also removed. The purpose of this study was to assess the safety and efficacy of 1 mg varenicline twice a day (bis in die, or “b.i.d,” the current FDA-approved full dose) in comparison with 0.5 mg b.i.d. varenicline and placebo without the advice to quit or oversight of a health care professional, meant to generalize to an OTC environment. In an early Phase 2 dosing ranging study, Nides, et al. concluded that the 0.5 mg b.i.d. dose was as efficacious as the dose of 1.0 mg b.i.d. but was associated with a 50% reduced incidence of nausea and abnormal dreams.¹⁸ Fewer side effects could lead to greater medication adherence, particularly in an OTC environment. We hypothesized that quit rates in 1 mg b.i.d. varenicline condition would be greater than placebo, but not greater than 0.5 mg b.i.d., and that 0.5 mg b.i.d. would be more effective than placebo. We further hypothesized that adverse events would show a dose–response, with overall adverse events lowest in the placebo group. To answer these questions, we employed a randomized, double-blind, placebo-controlled study to determine the safety and efficacy of switching smoking cessation medications from Rx-only to OTC.

Materials and Methods

Study Overview

This study was performed at two clinical sites in the United States (Burbank, California and Phoenix, Arizona). Participants were randomly assigned to one of three conditions: 1 mg varenicline b.i.d., 0.5 mg b.i.d. varenicline, and placebo b.i.d. Randomization was balanced within sites. In addition to the main study, 60 subjects (20 per condition) were randomized to participate in a 2-week ecological momentary assessment sub-study to examine how medication adherence, withdrawal, and side effects vary within subjects and are associated with study condition (e.g., 1 mg medication, 0.5 mg medication, or placebo) in the first 2 weeks of medication use (results not presented). The study protocol and patient-facing study forms were approved by the WIRB-Copernicus Group Institutional Review Board for both sites. An independent Data Safety and Monitoring Board reviewed adverse event data for patient safety.

Recruitment began in May 2018; new recruitment was halted from March 2020 to September 2020 due to the COVID-19 pandemic. Participants (n = 19) who were already enrolled in the study at the time of the pandemic were transitioned to telephone-only visits. Since they were not able to contribute biologically confirmed smoking abstinence data at all-time points, these participants were excluded from analyses. In January 2021, the protocol was modified (with IRB approval) to add a remote cohort with telephone-only assessments and biological confirmation of smoking abstinence using urine tests (Jant Accutest) to resume data collection while reducing COVID transmission risk; the remote cohort portion of the study was discontinued and the in-person portion of the study resumed in September 2021 for 2 weeks before a Pfizer voluntary recall of varenicline ended recruitment mid-September 2021.¹⁹ Enrolled participants were given the choice to continue in the study or discontinue. The final follow-up was completed in April 2022. Below, we highlight when aspects of the study protocol for the remote cohort differed from that of the in-person cohort.

Participant Recruitment

Participants were recruited using site databases, traditional media, flyers, social media posts, and an outside research company contracted to identify potentially eligible participants from their database.

Study Eligibility

To be eligible for the in-person cohort, participants had to be 21+ years of age, self-report daily smoking with a breath carbon monoxide (CO) ≥ 10 ppm, and be motivated to quit smoking within 5 weeks of the baseline visit (defined as >5 on a scale of 1–10 on reported motivation). Participants were excluded for the following reasons: self-reported diagnosis or treatment of serious cardiovascular disease or mental health condition within the last six months; current suicidal ideation (≥7 on the suicidal behavior questionnaire); self-reported diagnosis or treatment for depression within the past 6 months, unless participant has written permission by their health care provider to participate; systolic blood pressure of 160 or higher and/or diastolic blood pressure of 100 or higher; history of renal disease; allergy to any of the ingredients in varenicline; participation in another smoking cessation program or any type of clinical trial in the past three months; use of any smoking cessation medication in the past 3 months; positive drug screen indicating possible substance abuse (e.g. opiates, amphetamines, benzodiazepines, cocaine, or other substances, but not cannabis), unless participant can show that the medication has been prescribed by licensed clinical provider; consume greater than 21 alcohol drinks per week; or females who are pregnant, breast feeding, or not currently using a medically approved form of birth control and unwilling to do so.

Eligibility requirements for the remote cohort were the same, except (1) the exhaled CO was replaced by a positive result from a mailed urine cotinine test (Jant Accutest; ≥200 ng /mL; participants took a picture of the results and sent them to study staff to confirm eligibility); and (2) use of illicit drugs or nonprescription medication use was assessed via self-report during screening.

While these inclusion criteria are somewhat restrictive, it is important to note that OTC smoking cessation studies designed to meet FDA criteria, such as those investigating nicotine replacement,¹⁴ were not “real world” studies that allowed anyone to obtain the medication but included inclusion and exclusion criterion similar to non-OTC studies. FDA requires controlled studies that do not include an evaluation by a prescribing health care provider to determine whether a prescription medication can be switched to OTC, and the present study was designed to satisfy this requirement. Indeed, the authors sought guidance from the FDA before implementing the study.

Study Procedures

Study Protocol Summary

Potential participants were phone screened for eligibility and were scheduled for a baseline visit if eligible. At baseline, participants were provided with time to read a written informed consent form, reviewed the form with study personnel, and were encouraged to ask questions. The informed consent process satisfied standards according to Good Clinical Practice (GCP), the International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use (ICH), and FDA guidelines. After consent, participants were rescreened to ensure all-inclusion/exclusion criteria were still met (e.g., smoking status via breath CO in the in-person visits and via urine cotinine in virtual visits), randomized into one of the three treatment conditions, and given the study medication to begin taking the following day. In-person cohort study visits were scheduled for weeks 2, 4, 8, and 12; study phone calls were scheduled for weeks 13 and 26. Medication use, adverse events, withdrawal symptoms, and smoking status were assessed at each visit.

Compensation

If they participated in all visits, main study participants could be compensated up to $300 pre-COVID pause; this amount was increased to $350 to increase recruitment when the study restarted in January 2021.

Research Product

Varenicline (1 mg and 0.5 mg) and placebo was provided by Pfizer at no cost; Pfizer had no other role in the study, does not own the data, and had no control over manuscript content, though the draft of this manuscript was shared with Pfizer pre-publication per the terms of the medication agreement. Due to pill coloration differences based on dosage, we encapsulated the medication, including the placebo, in identical capsules to ensure uniformity and to protect against blinding under the direction of a research pharmacist, who implemented a protocol to ensure product safety, proper blinding, and correct storage. Pills were provided to participants in pill bottles.

Intervention

Participants in the treatment conditions received a placebo, 0.5 mg tablets of varenicline b.i.d. or 1 mg tablets of varenicline b.i.d. in a double-blind fashion. The in-person cohort participants received their medication from study staff during study visits; remote cohort participants were sent their medications. The medication dosing explanation and schedule followed the Pfizer medication instructions. To maintain the blinding, all-participants took one capsule in the morning for the first 4 days of treatment, then took two capsules per day, one in the morning and one in the evening, for the remainder of treatment. Participants in the 0.5 mg b.i.d. condition took 0.5 mg in the morning for the first four days. On day five, they began the full dose of 0.5 mg in the morning and evening for the remaining 12-week treatment period. Participants in the 1 mg condition took 0.5 mg once per day, then took 0.5 mg twice per day until day eight when they began taking 1 mg twice per day for the remainder of the study. All-conditions took the pills for 12 weeks. Study staff provided participants with the manufacturer’s package insert on how to use the medication, including recommendations on how to pick a quit date, though participants were not required to formally commit to an exact date. Participants did not receive any form of counseling or behavioral support from the study staff throughout the treatment period. The only behavioral support offered was information on the label encouraging participants to call the national smoking cessation quitline (800-QUITNOW).

Measures—Primary Outcome

Biologically Confirmed Continuous Smoking Abstinence

The primary outcome was biologically confirmed continuous smoking abstinence between weeks 8 and 12, defined as no self-reported smoking between weeks 8 and 12, with biochemical verification via exhaled CO of <10 ppm or urine cotinine test (for the remote cohort) at week 12. Additional efficacy outcomes were biologically confirmed continuous smoking abstinence between weeks 8 and 26 and 12 and 26, defined similarly.

The method for collecting biologically confirmed smoking abstinence varied by cohort. In-person cohort participants self-reported smoking status at week 8 and 12 clinic visits; those who reported not smoking in the past seven days provided exhaled CO (Bedfont Micro+) to objectively confirm self-report (<10 ppm). Participants also self-reported their smoking status at a follow-up phone call at week 26; in-person participants who reported not smoking within seven days of the week 26 phone call were asked to come into the clinic to provide an exhaled CO, while remote participants completed urine cotinine tests (Jant Accutest) and emailed pictures of the urine cotinine results to the study staff for verification.

Measures—Secondary Outcomes

Biologically Confirmed Point Prevalence Smoking Abstinence

We also report the point prevalence of smoking abstinence at weeks 2, 4, 8,12, and 26. Point prevalence smoking abstinence was defined as no self-reported smoking in the past 7 days, with biochemical confirmation via biological confirmation (CO or urine cotinine test) at that week’s visit.

Withdrawal

We also report changes in withdrawal as measured by the Minnesota Tobacco Withdrawal Scale (MTWS), a common and robust measure of tobacco withdrawal.^20–24

Adverse Events

Evaluation of adverse events across treatment conditions was included to assess the relative safety of varenicline when no health care provider was involved. Adverse events were elicited from participants at every study visit starting at visit 2.

Adherence

For the main study, medication adherence was defined as taking at least 80% of the prescribed pills. Participants brought in their remaining pills (for the in-person cohort) or self-reported the number of remaining pills (remote cohort) during the weeks 2–12 visits. Participants who were missing pill count data were assumed to have taken zero pills preceding that visit.

Data Analysis Plan

Missing Data

As is recommended best practice,²⁵ we assumed that participants who were lost to follow-up continued to smoke. We also assumed that participants who did not provide biochemical verification of their self-reported smoking abstinence would have failed their biochemical verification. We made no assumptions about participants’ MTWS or adverse event data if they were lost to follow-up.

Main Study Analyses

Our a priori primary hypotheses were that 1 mg b.i.d. and 0.5 mg b.i.d. varenicline would be more effective than placebo and that there would be no efficacy difference between 1 mg b.i.d. and 0.5 mg b.i.d. varenicline, as measured by biochemically verified continuous abstinence between weeks 8 and 12 (end of medication use).

We used generalized linear mixed models (GLMM) to determine if study mode (in-person, remote, and hybrid) or site (Arizona vs. California) was related to quit outcomes. The purpose of this preliminary analysis was to determine if the main hypothesis tests needed to be broken out by cohort location or site. Next, we used GLMM to evaluate differences in continuous smoking abstinence by condition and week and the interaction of condition and week. All-time points were included in these models and the participant was included as the random effect to account for dependencies created by the repeated measures. We then used logistic regressions to determine was there was a relationship between study condition and point prevalence abstinence, adverse event reporting, or adherence to study medication. Statistical significance was determined by p values < .05. Main study analyses were conducted in SAS 9.4 with the GLIMMIX and LOGISTIC procedures.

Results

CONSORT Diagram

Three-hundred and eighty-three individuals were screened and 313 met the screening criteria and were randomized across the two sites. Participants were randomly assigned to one of three conditions: 104 were assigned to receive 1 mg varenicline b.i.d., 104 were assigned to receive 0.5 mg b.i.d. varenicline, and 105 participants were assigned to receive placebo b.i.d. Pre-randomization reasons for exclusion and post-randomization reasons for not completing the study are provided in the CONSORT diagram (Figure 1).

Figure 1.

CONSORT diagram.

Baseline Characteristics

Baseline demographic and smoking characteristics by study condition are presented in Table 1. Overall, the study population was 38.0% men, 57.8% Caucasian, and 88.2% non-Hispanic; the mean age was 51.5 years (SD 11.3). Most participants (39.9%) had completed some college but did not have a college degree. Participants smoked on average 16–17 cigarettes per day, had a Fagerstrom score of 4 (SD 1.4; out of a maximum of 10) and an MTWS score of 10.8 (SD 7.3; out of a maximum of 60). There were no differences in demographic or smoking-related variables by condition.

Table 1.

Demographic and Smoking Behavior Sample Characteristics (n = 313)

		Condition
	Total (n = 313)	Placebo (n = 105)	0.5 mg varenicline (n = 104)	1 mg varenicline (n = 104)	p-Valuea
Age (mean, SD)	51.5 (11.3)	51.0 (12.3)	51.4 (11.2)	52.0 (10.5)	.21
Sex (n (%))b					.17
Male	119 (38.0%)	33 (31.4%)	39 (37.5%)	47 (45.2%)
Female	193 (61.7%)	72 (68.6%)	65 (62.5%)	56 (53.9%)
Missing	1 (0.3%)	–	–	1 (0.9%)
Race (n (%))b					.18
White or Caucasian	181 (57.8%)	63 (60.0%)	58 (55.8%)	60 (57.7%)
Black or African American	119 (38.0%)	35 (33.3%)	44 (42.3%)	40 (38.5%)
Other small groupsc	7 (2.2%)	5 (4.8%)	2 (1.9%)	–
Biracial	5 (1.6%)	2 (1.9%)	–	3 (2.9%)
Missing	1 (0.3%)	–	–	1 (1.0%)
Ethnicity (n (%))b					.71
Non-Hispanic	276 (88.2%)	92 (29.4%)	92 (29.4%)	92 (29.4%)
Hispanic	36 (11.5%)	13 (4.2%)	12 (3.8%)	11 (3.5%)
Missing	1 (0.3%)	–	–	1 (0.3%)
Education (n (%))b					.26
Some high school or less	19 (6.1%)	9 (2.9%)	3 (1.0%)	7 (2.2%)
High school graduate or GED	68 (21.7%)	21 (6.7%)	24 (7.7%)	23 (7.4%)
Some college, no degree	125 (39.9%)	44 (14.1%)	45 (14.4%)	36 (11.5%)
Associate’s degree	41 (13.1%)	9 (2.9%)	17 (5.4%)	15 (4.8%)
Bachelor’s degree	50 (16.0%)	20 (6.4%)	14 (4.5%)	16 (5.1%)
Graduate degree	9 (2.9%)	2 (0.6%)	1 (0.3%)	6 (1.9%)
Missing	1 (0.3%)	–	–	1 (0.3%)
Cigarettes per day,  n2 (mean, SD)	n = 310 16.8 (8.4)	n = 105 17.2 (7.7)	n = 102 16.4 (8.2)	n = 103 16.7 (9.3)	.76
Fagerstrom test for nicotine Dependence Score,  n2 (mean, SD)	n = 309 4.0 (1.4)	n = 105 4.1 (1.4)	n = 102 3.8 (1.4)	n = 102 4.1 (1.3)	.25
Minnesota tobacco withdrawal scale,  n2 (mean, SD)	n = 307 10.8 (7.3)	n = 103 10.7 (6.7)	n = 102 10.5 (7.4)	n = 102 11.3 (7.6)	.97

^aTest for between-condition differences using chi-square for categorical variables and ANOVA for continuous variables.

^bCount was less than the total due to participants being randomized but withdrawing consent prior to the continuation of data collection.

^cCategories include Asian, Native Hawaiian or other Pacific Islander, American Indian or Alaska Native, and “Other”.

Preliminary Analyses of Study Location and Mode

There was no significant difference in biochemically confirmed continuous quit rate by study site (F = 2.88, p = .09). There was also no significant interaction between study mode and study site (F = 0.15, p = .86). Thus, participants did not quit at different rates by study mode or site and there is empirical evidence to justify analyzing the data by grouping the participants from different study modes and sites together.

Primary Outcome—Biologically Confirmed Continuous Abstinence Quit Rate

The proportion of participants with biologically verified continuous abstinence at weeks 4, 8, 12, and 26 is presented in Figure 2. The interaction between condition and week was not significant (weeks 8–26: F = 0.29, p = .75; weeks 12–26: F = 0.39, p = .68). Odds ratios for comparisons by condition and outcome, after Bonferroni adjustments for alpha inflation, are presented in Table 2.

Figure 2.

Proportion of participants with biologically confirmed continuous smoking abstinence at weeks 4, 8, 12, and 26.

Table 2.

Odds Ratios (95% CI) for Group Comparisons Between Weeks by Outcome

	Condition OR (95% CI)
Comparison	0.5 mg b.i.d vs. Placebo	1.0 mg b.i.d. vs. Placebo	1.0 mg b.i.d. vs. 0.5 mg b.i.d.
Continuous abstinencea
8 vs. 12 weeks	1.18 (0.39, 3.61)	1.99 (0.70, 5.63)	1.68 (0.61, 4.68)
8 vs. 26 weeks	1.02 (0.37, 2.80)	1.42 (0.55, 3.68)	1.39 (0.53, 3.61)
12 vs. 26 weeks	0.90 (0.32, 2.48)	1.64 (0.65, 4.18)	1.83 (0.70, 4.80)
Point prevalence abstinenceb
8 weeks	1.14 (0.48, 2.73)	2.03 (0.91, 4.50)	1.77 (0.81, 3.87)
12 weeks	1.43 (0.57, 3.57)	3.39 (1.49, 7.71)	2.37 (1.11, 5.05)
26 weeks	0.84 (0.35, 2.06)	1.30 (0.57, 2.93)	1.54 (0.65, 3.61)

^aContinuous abstinence results are from a generalized linear mixed model.

^bPoint prevalence abstinence results are from logistic regression.

Secondary Outcome—Biologically Confirmed Point Prevalence Quit Rate

The proportion of participants reporting biochemically confirmed point prevalence smoking abstinence at weeks 2, 4, 8, 12, and 26 is presented in Figure S1. The odds of biochemically confirmed point prevalence abstinence were significantly higher for those assigned to the 1 mg b.i.d. condition than for those assigned to the placebo condition at week 12 (OR 3.39; 95% CI 1.49, 7.71). The odds of biochemically confirmed point prevalence abstinence were also significantly higher for those assigned to the 1.0 mg b.i.d. condition than the 0.5 mg b.i.d. condition at week 12 (OR 2.37; 95% CI 1.11, 5.05; Table 2).

Secondary Outcomes—MTWS Score

The mean MTWS score by condition and week is presented in Table S1. Compared to placebo, there was no significant difference in MTWS score for the 1.0 mg b.i.d condition between weeks 4 and 8 (F = 0.18, p = .67) and between weeks 8 and 12 (F = 1.83, p = .18). Compared to placebo, there was no significant difference in MTWS score for the 0.5 mg b.i.d group between weeks 4 and 8 (F = 0.03, p = .87) and between weeks 8 and 12, (F = 0.00, p = .96). There were also no differences in MTWS score comparing the 1.0 mg to the 0.5 mg b.i.d group between weeks 4 and 8 (F = 0.05, p = .83) and between weeks 8 and 12, (F = 1.23, p = .27).

Secondary Outcomes—Adverse Events

Adverse event rates are presented in Table S2. In the placebo condition, 51 adverse events were reported (AE rate of 48.6%); 58 adverse events were reported in the 0.5 mg varenicline condition (AE rate of 56.3%); and 58 events were reported in the 1 mg varenicline condition (AE rate of 56.9%). There was no difference in adverse events by treatment condition comparing the 0.5 mg b.i.d. to placebo (OR 0.72; 95% CI .42, 1.26) or 1.0 mg vs. placebo (OR 0.72 (0.41, 1.24)), nor was there a difference when comparing 1.0 mg b.i.d. to 0.5 mg b.i.d. (OR 0.98; 95% CI .56, 1.70). Consistent with other studies, the two most frequently reported adverse events by treatment condition were nausea (11.8% in placebo, 31.0% in 0.5 mg b.i.d., and 36.2% in 1 mg b.i.d.) and vivid dreams (15.7% in placebo, 20.7% in 0.5 mg b.i.d., and 22.4% in 1 mg b.i.d.).

Secondary Outcomes—Adherence

Overall, 41% (43 participants) were adherent in the placebo condition, 39.8% (41 participants) were adherent in the 0.5 mg b.i.d. condition, and 39.2% (40 participants) were adherent in the 1.0 mg b.i.d. condition. There were no differences between study conditions regarding adherence when comparing the 0.5 mg b.i.d. condition to placebo (OR 1.05; 95% CI .60, 1.83), the 1.0 mg b.i.d. to placebo (OR 1.08; 05% CO 0.62, 1.88), or the 1.0 mg b.i.d. to the 0.5 mg b.i.d. (OR 1.03; 95% CI 0.59, 1.80).

Discussion

Our main hypothesis, that 1 mg and 0.5 mg b.i.d. varenicline would result in higher biologically confirmed continuous abstinence quit rates at the end of treatment than placebo, was not supported. However, point prevalence quit rates in the 1 mg b.i.d. varenicline condition were significantly higher than 0.5 mg b.i.d. varenicline and placebo at the end of treatment (week 12). The trends in continuous abstinence rates were consistent with those of the point prevalence outcomes, suggesting that 1 mg b.i.d. resulted in higher quit rates than both 0.5 mg b.i.d. and placebo. Unfortunately, the COVID pandemic and later medication recall prevented the recruitment of the planned sample of 405 participants, so the study was underpowered to detect significant differences when assessing continuous abstinence. Nonetheless, even with an underpowered study, the consistent direction of results suggests that 1 mg b.i.d. varenicline used in an OTC environment could help people who smoke quit and that the medication has the potential to switch to OTC status.²⁶ This could make it more readily accessible to far more people who smoke who want to quit, particularly those who will not or cannot go to a health care provider, and perhaps could benefit those who are not committed to quitting but might want to try the medication when they see it available in a pharmacy.²⁷ However, additional research is needed given the uneven results of this study, the extensive exclusion criteria used to gather the sample, which limits generalizability, and the prescribed nature of the dosing, which may not approximate how people take varenicline in a real-world OTC environment.

Because this study did not provide any behavioral support, our results showing somewhat lower quit rates (17.0% continuous abstinence rates at week 12 relative to the 29% found in EAGLES)¹¹ is not surprising. Carpenter et al¹⁵ and Swan et al¹⁴ found 12-week self-reported point prevalence quit rates of 18% and 39%, respectively, in 1 mg b.i.d participants, while our 12-week biochemically verified point prevalence quit rates were 25.0% at week 12. Our results relative to EAGLES may also be explained by our adherence rate, which was only half that of what was reported in the EAGLES psychiatric cohort, a likely underestimation of the overall EAGLES adherence rate.²⁸ Our results were consistent with the Cochrane meta-analyses on the efficacy of varenicline for smoking cessation, which consisted of studies that typically included behavioral support and health care provider assessment.²⁹ Added to that support is our finding that adverse events from medication use were modest and manageable; contrary to our hypothesis, we did not see a significant difference in adverse events by condition. Thus, the results of this study partially support the premise that varenicline can be effective as an OTC medication.

One concern is that quit rates in the 1 mg b.i.d. condition decreased to a level similar to the other study conditions after medication use was completed. This has been observed in research on real-world use of varenicline and reflects the need for additional research on longer-term use of varenicline as an OTC and exploration of ways to minimize relapse when medication use has ended in an OTC environment, as prior studies have demonstrated that longer-term use of varenicline increases abstinence.³⁰ Moreover, the overall quit rates were lower than those in studies where behavior support was provided, which is consistent with studies on the switch to OTC nicotine replacement.²⁵ Thus, even though the results of this study suggest that varenicline may help people who smoke quit with modest side effects, the best approach to help people who smoke quit is varenicline in combination with behavior support.

This study has multiple limitations mainly due to the circumstances in which it was conducted. In addition to our inability to reach our recruitment goal due to COVID and a medication recall, the COVID pandemic also prevented in-person visits. Thus, we were required to shift the completion of the study by allowing remote inclusion in the study. This change made it more challenging to verify abstinence, but it also allowed the implementation of new approaches to implementation of clinical smoking cessation research that will likely inform changes in clinical trials in the future. Despite including both in-person and remote participation, we found that combining these two groups did not change the outcome of the study. Additional limitations include limited baseline information about the participants, particularly concerning their prior varenicline use and quit history.

This study points to multiple areas of needed research on varenicline, including evaluation of OTC varenicline study in the “real world” by assessing outcomes in a nonclinical environment so that effectiveness and not just efficacy can be assessed,³¹ as well as evaluation of OTC varenicline with a less restricted population of participants than employed in this study. With increased emphasis by the US government on helping more people quit smoking³² and increasing interest in policy approaches to decreasing smoking prevalence such as banning menthol cigarettes, switching varenicline to an OTC medication could have significant public health benefits.

Supplementary material

Supplementary material is available at Nicotine and Tobacco Research online.

Funding

This grant was funded by the National Institutes of Health, National Institute on Drug Abuse grant number R01 DA044125. Study medication was provided at no cost by Pfizer grant # WI234549.

Declaration of Interests

SJL received medication for this study at no cost from Pfizer and received funding from Achieve Life Sciences to conduct a clinical trial on cytisinicline (cytisine) for smoking cessation. MAN has received consulting fees from Achieve Life Sciences, a manufacturer of cytisinicline (cytisine) for smoking cessation and has conducted PK/abuse liability studies for manufacturers of nicotine-containing products including Electronic Nicotine Delivery Systems and nicotine pouches. JP is a paid expert witness for the Plaintiffs in a Multi-District Litigation invoking Natural American Spirit Cigarettes and a paid expert witness for the Plaintiffs in a Derivative Litigation invoking Altria Group, Inc. This arrangement has been reviewed and approved by the University of Nevada, Reno in accordance with their conflicts of interest policies. The following coauthors have no declared conflicts: MB, RS, JK, MM, JM, and AC.

Author contributions

Jennifer Pearson (Conceptualization [supporting], Formal analysis [equal], Funding acquisition [supporting], Investigation [supporting], Methodology [equal], Project administration [equal], Resources [supporting], Writing—original draft [supporting], Writing—review & editing [supporting]), Mitchell Nides (Conceptualization [supporting], Funding acquisition [supporting], Investigation [equal], Methodology [equal], Project administration [equal], Resources [equal], Supervision [equal], Writing—original draft [supporting], Writing—review & editing [supporting]), Alana Cataldo (Investigation [supporting], Methodology [supporting], Project administration [equal], Supervision [supporting], Writing—original draft [supporting], Writing—review & editing [supporting]), Marisol Martinez (Investigation [supporting], Project administration [supporting], Supervision [supporting]), Jackie Morales (Investigation [supporting], Project administration [supporting], Writing—review & editing [supporting]), Ryan Seltzer (Data curation [equal], Formal analysis [lead], Investigation [supporting], Project administration [supporting], Validation [Equal], Writing—original draft [supporting], Writing—review & editing [supporting]), Jonathan Kurka (Data curation [equal], Methodology [supporting], Project administration [supporting], Software [equal], Validation [equal], Writing—review & editing [supporting]), Matthew Broussard (Project administration [supporting], Writing—original draft [supporting], Writing—review & editing [supporting]), and Scott Leischow (Conceptualization [lead], Funding acquisition [lead], Investigation [lead], Methodology [equal], Project administration [lead], Resources [lead], Supervision [lead], Writing—original draft [lead], Writing—review & editing [lead])

Data availability

The data underlying this article are available at https://doi.org/10.48349/ASU/DYPIMH.