A Pilot Randomized Clinical Trial of Remote Varenicline Sampling to Promote Treatment Engagement and Smoking Cessation

Abstract

Introduction

Medication sampling is a clinically useful tool to engage smokers in the quitting process. Whether varenicline is suitable for sampling purposes is unclear. The purpose of this study was to examine the feasibility, uptake, and preliminary outcomes of varenicline sampling.

Methods

Smokers (N = 99), both motivated to quit and not, were recruited and randomized to varenicline sampling versus not, with 12 week follow-up. The intervention consisted of mailing one-time samples of varenicline (lasting 2–4 wks), with minimally suggestive guidance on use.

Results

Uptake of varenicline was strong, at 2 weeks (54% any use, 66% daily use) and 4 weeks (38%, 46%), with 58% of medication users seeking additional medication. Most users followed conventional titration patterns, self-titrating from 0.5 mg to 2 mg. Relative to control, varenicline sampling increased motivation (p = 0.006) and confidence to quit (p = 0.02), and decreased cigarette smoking (p = 0.02). Smokers receiving varenicline samples were significantly more likely to achieve 50% reduction in cigarettes per day (CPD), both immediately following the sampling exercise (Adjusted Odds Ratio [AOR] = 4.12; 95% CI: 1.39 to 12.17) and at final follow-up (AOR = 4.50; 95% CI: 1.56 to 13.01). Though cessation outcomes were not statistically significant, there was a 1.5 to 3-fold increase in quit attempts and abstinence from varenicline sampling throughout follow-up. These outcomes were comparable among smokers motivated to quit and not.

Conclusions

Unguided, user-driven sampling of varenicline sampling is a concrete behavioral exercise that is feasible to do and seems to suggest clinical utility. Sampling is a pragmatic clinical approach to engage more smokers in quitting.

Implications

Use of evidence-based pharmacotherapies for smoking cessation is low. Medication sampling is a pragmatic behavioral exercise that allows smokers to experience the benefits of using them, while promoting positive downstream effects towards quitting. While previous studies have shown that nicotine replacement therapy (NRT) sampling is viable and effective, whether this extends to varenicline is unclear. Results from this trial demonstrate that varenicline sampling is feasible, safe, and suggestive of clinically important steps toward quitting, deserving of a larger trial.

Clinical Trial Registration

NCT #03742154.

Background

USPHS clinical practice guidelines¹ and other meta-analyses^2–4 clearly recommend that (1) all tobacco users should be offered brief advice to assist in quitting, and (2) unless medically contraindicated, all smokers should be offered pharmacotherapy. Our studies^5,6 and those of many others^7–11 consistently demonstrate that use of pharmacotherapies is unacceptably low, with <25% of quit-attempters using them. There is strong public health need to capitalize on existing treatments and deliver them in ways that specifically enhance smokers' capacity to initiate and sustain evidence-based quit attempts. Such interventions should be easily accessible and offer potential for scalability into larger populations of smokers. Medication sampling may meet this need.

Medication sampling simply refers to providing starter packs of cessation medications. Unlike traditional models that first rely on smoker intent to quit, sampling is readily scalable to smokers regardless of their quit intention, offering a concrete, behavioral cue to facilitate quitting. It is not meant to replace more intensive treatment but rather serve as a catalyst to it. In prior trials of nicotine replacement therapy (NRT) sampling, we^12–16 and others^17–19 have shown that sampling can engage smokers in use of evidence-based treatment, induce quit attempts and cessation. Aside from pharmacological effects of reduced craving, sampling NRT can increase motivation, confidence and familiarization/acceptance of medications.²⁰

Varenicline is arguably the most effective single agent medication for smoking cessation,^21–23 and is a core component of clinical practice guidelines.^24–26 Whether varenicline fits within a sampling paradigm is a compelling question. Smokers endorse strong preference for varenicline over other options.²⁷ Several studies have examined nontraditional delivery of varenicline, that is, outside the context of a cessation-focused clinical trial. This includes studies that provided varenicline without any direct clinical interview,^28–30 through mail,^28–30 to unmotivated smokers,^31–33 with lower dosing,^34,35 for short duration,²⁸ or allowing for self-paced use/flexible quit dates.^32,36 Collectively, these findings would suggest that unguided, smoker-driven sampling of varenicline is viable and worth testing. On the other hand, there are certainly a range of reasons why varenicline might not be effective. Unlike other cessation medications, varenicline has a standard titration schedule of escalating dose, requires consistent scheduled dosing, requires a prescription, and may produce adverse events requiring symptom management, all of which might require more hands on, not less.

We herein report on a pilot remote clinical trial of varenicline sampling versus not, focusing on feasibility, uptake and safety of varenicline sampling, providing descriptive outcomes of each. We were secondarily focused on between-group differences (clinical if not statistical) on: (1) quit attempts, (2) abstinence, (3) smoking reduction, as well as (4) hypothesized mediators (confidence, motivation to quit [MTQ], attitudes towards medication).

Methods

Trial Overview

A demographically diverse sample of 99 smokers across South Carolina was consented and randomized to receive a 4-week mailed sample of varenicline, or not. Recruitment included purposeful balance of low and high MTQ, allowing for exploratory analyses within each motivational subgroup. As sampling is largely a behavioral exercise inclusive of medication expectancies, there were no study blinds. Consistent with our prior sampling studies, our “intervention” included suggestive instructions on use, consistent with what might be briefly provided within a physician's office. However, messaging was kept minimal, and self-determined choice of use was emphasized. The trial began on 11/2018 and concluded on 4/2020. All procedures were approved by IRB, and the trial was publicly registered (NCT# 03742154).

Participants

Eligibility was contingent upon: (1) age 18+; (2) daily smoker (25+ days/month); (3) smoking 5+ cigarettes/day; (4) smoking ≥1 year; (5) some eventual interest in quitting smoking (≥2 on 0–10 motivation ladder); (6) having a primary care doctor, seen ≥1 visit in past year; (7) not currently pregnant, breastfeeding, or planning to become pregnant (see below for remote pregnancy verification); (8) no suicidal ideation in past month, nor any lifetime suicide attempt; (9) no current (past month) psychotic symptoms (auditory/visual hallucinations) nor self-reported lifetime diagnosis of or treatment for bipolar disorder, schizophrenia, or other psychotic disorders; (10) no medical contraindications for varenicline (eg cardiac or renal disease), and (11) own a smartphone (either iOS or Android), or have regular email access.

Our rationale for having an established primary care provider was to allow for and promote conversion; so that smokers could readily obtain varenicline after sampling it, if desired. We required smartphone/email access to facilitate access to follow-up research assessments.

Recruitment, Teleconsent, Baseline Assessment, and Randomization

Initial Screening

Advertising was based largely within online recruitment platforms (eg Craigslist) wherein potential enrollees linked to a screener to determine study eligibility (Figure 1). This online screener assessed all eligibility criteria above, with the exception of mental health assessment and pregnancy verification, which were collected after consent as a part of secondary screening (see below). Study enrollment was a priori balanced on low (0–6 on 0–10 ladder; n = 50) versus high (7–10; n = 50) MTQ in the next month. Thus, some individuals were administratively excluded (see Figure 1) if either of these motivational “bins” were filled.

Figure 1.

CONSORT flow.

Consent

Eligible participants were consented through one of two options. The first was via tele-consent which included a real-time synchronous discussion and exchange of signed documents between research staff and potential enrollee, culminating in an electronically signed, shared and mutually kept document. The second consent option, for those who preferred or who did not have tele-consent ability, was based on mailing of consent paperwork and self-addressed, pre-stamped return envelope. Upon returned receipt of the signed consent to the central research office, research staff engaged in a phone-based discussion of the consent form and study procedures. The proportion of tele- versus mailed consent was 65% versus 35%, respectively. All consented participants then underwent secondary screening.

Secondary Screening

The purpose of secondary screening was to ascertain clinical eligibility, excluding anyone with suicidal history (n = 1) or anything else that might have changed since initial screening (see Figure 1). Questions of past month suicidal ideation were asked post-consent, after contact information was collected in case of clinical emergency. Past month clinical symptomatology (criterion 8) were based on both the NIMH toolkit for suicidal ideation (5 yes/no items), as well as brief items from the MINI.³⁷ In addition, secondary screening collected verification of pregnancy status. All women of child-bearing age (≤55 years old) were sent a separate mailed packet for remote pregnancy testing. Upon receipt and self-testing, these individuals were sent a participant-specific url and code, and attested via electronic signature to their negative (non-pregnant) result. Women could not go forward in the study unless this attestation was received.

Baseline Assessment

Following all steps above, participants completed a brief assessment that included demographic and smoking history questions. The latter included assessment of cigarette smoking, dependence,³⁸ quit history and prior use of cessation strategies. Within demographic history, we categorized rurality (as a potentially relevant measure of health disparities) based on zip code, following prior precedent.^39,40 Following baseline assessment, participants were then randomized and received the one-time intervention (sampling vs. not) at Baseline/Week 0.

Randomization

Enrollees were randomized in a 1/1 ratio to sampling/control groups using a mixed block design stratified by MTQ in the next month. As noted, we purposefully recruited equal numbers of smokers wanting to quit (≥7 on 0–10 scale) and not to ensure balance between groups.

Interventions and Messaging

At baseline all participants received oral and written suggestion to quit, including a recommendation (contact information and benefits) for use of the South Carolina quitline. Those in the sampling group also received a supply of varenicline in a childproof pill bottle, with no other intervention beyond this mailing. During the baseline call and within this mailing, guidance on varenicline was kept minimal, merely suggesting possible use and potential benefits of use (see Appendix A for actual mailing insert, with visual depiction of potential titration). In addition, varenicline participants received a letter that they could give to their healthcare provider at their own discretion, as an attempt to facilitate continued use if they wished.

Typical practice is to titrate varenicline as follows: 0.5 mg QD for days 1–3, 0.5 mg BID for days 4–7, and 1 mg BID from day 8 and beyond. On one hand, we wanted to provide the strongest, most evidence-based dosage possible. If varenicline sampling resulted in a null effect, we did not want this to be because of dose. On the other hand, had we provided multiple dosages within a self-paced sampling experience this could have led to unwanted confusion and risk. Two large trials showed similar outcomes of 1 mg versus 2 mg dosing, with fewer adverse events in the former and somewhat stronger effects in the latter.^34,35 To balance these perspectives, we: (1) provided 0.5 mg tablets only, 56 total (enough for 30 days at 1 mg or 14 days at 2 mg) with instructions to titrate as noted in the mailed information sheet, (2) gave instructions to take no more than two in the morning and two in the evening, and (3) gave the participant the ultimate choice to self-pace usage. Thus, we viewed the sampling experience as lasting 2–4 weeks depending on participants' choice.

Follow-up Procedures and Assessments

Participants were auto-sent a survey link at +17 days (2 weeks beyond initial mailing, accounting for ~3 day delay in mailing), +33 days (same delay), +8 weeks, and +12 weeks. Participants were reimbursed for completion of each of five assessment points at an escalating rate, up to $160 total. Uptake of varenicline, within the sampling group alone, was defined as (1) any use, (2) regular/daily use, and (3) independent use (using varenicline on their own, beyond what was provided). As this was a naturalistic sampling experience, we were also interested in self-chosen titration patterns, that is, patterns of self-dosage of over time (0–2 mg, in 0.5 mg increments). All participants, including control group, were asked about general medication use (varenicline, NRT, other) throughout follow-up. Motivation and confidence to quit were assessed using single item measures (0–10 visual analog scale) for each. Attitudes towards varenicline were assessed via: (1) Likert-scaled questions that asked about varenicline safety and efficacy (1 item each), and (2) an adapted 6-item questionnaire that has been previously published for beliefs and attitudes toward bupropion⁴¹ (switched to reference varenicline instead; referred to as “BAV” below). Using Timeline Follow Back (TLFB) methods, we assessed all smoking outcomes, including: cigarettes per day (CPD), averaged over the previous seven days, and from this we calculated abstinence, smoking reduction, and percent who achieved 50% reduction. Abstinence was defined as both “floating” (any seven day period of non-smoking) and point prevalence, as recommended for “cessation-induction” studies.⁴² The incidence of self-reported and 24-h quit attempts (both based on intention to quit for good) was defined as occurring at any point during over the 3-month follow-up period. Adverse events were tracked and rated as mild, moderate or severe by the patient and rated for relatedness to varenicline. In addition, our procedures included automated tracking of any instance in which participants reported significant increases in depressive symptoms (defined as an increase of ≥5 points since baseline with a resulting score ≥10), at any follow-up, as reported in the Patient Health Questionnaire/PHQ-9),⁴³ and also any endorsement of suicidal ideation (PHQ item #9). These procedures were added in large part because of the remote context of the trial. Any instance of either were clinically managed within 24 h through a direct phone call from a licensed clinical psychologist (MJC or JD). These outcomes are reported below, but in no instance did a participant report imminent risk for self-harm.

Data Analyses

The overarching goals of this study were focused on feasibility and uptake parameters around varenicline sampling. Sample size was largely driven by constraints in funding support, and thus between-group comparisons were likely underpowered. We were specifically interested in whether sampling effects generalized to those who are and are not motivated to quit, and did so via subgroup analyses. Using conservative intent-to-treat principles, any participant missing at follow-up assessment was assumed not to have made any quit attempts, nor quit. Baseline variables were compared between groups via two-sample t-tests, Chi-squared tests, or Fisher's exact tests, as appropriate. Variables that were different between the groups were considered as covariates in modeling of outcomes. Binary outcomes (incidence of quit attempts; 50% reduction in cigarettes per day; % treatment uptake/conversion, etc) were first examined over the entire study period from baseline to end of follow-up by calculating frequencies/percentages as well as 95% confidence intervals for each group, with chi-square tests (or Fisher's exact tests as needed) for between-group comparisons. All outcomes, both binary (over the entire study) and continuous outcomes measured at multiple time points during the study were evaluated using generalized linear mixed modeling with a main effect of group to determine group differences. For continuous outcomes, models included a main effect of time as well as a group by time interaction and a random subject effect to account for repeated measures within individuals. All models were adjusted for baseline covariates. Adjusted odds ratios (AOR) and 95% confidence intervals (95% CI) are reported on the main effect of treatment group (varenicline sampling vs. control) from these mixed models. Some outcomes were relevant to the varenicline sampling group only and are reported descriptively only.

Results

Sample Characteristics

Participants were slightly over 40 years of age, predominantly White, with at least some college educational history, but with moderately low household income (Table 1). There was an unexpected imbalance in gender across groups. While most participants had prior history with NRT, very few (9%) had ever used varenicline previously. Cigarette consumption, and by extension cigarette dependence (HSI), was slightly higher among those in the varenicline group. Aside from gender, CPD, and HSI, there were no other baseline differences between varenicline and control groups. Thus, for all analyses below, gender and HSI are included as covariates for all between-group comparisons (unadjusted results were comparable throughout). There were no differences in retention rates across follow-up (Figure 1).

Table 1.

Baseline Characteristics

	Varenicline (n = 50)	Control (n = 49)	p value
Demographics
Age (mean, SD)	43.0 (12.8)	40.4 (11.5)	0.3
Gender (n, % male)	28 (56%)	14 (29%)	0.006
Race			0.7
White	40 (80%)	35 (71.4%)
Black	8 (16%)	10 (20%)
Multiracial	2 (4%)	3 (6%)
Household income			0.8
<$25K	16 (32%)	15 (31%)
$25–$50K	22 (44%)	20 (41%)
$50–$75K	5 (10%)	6 (12%)
>$75K	3 (6%)	7 (14%)
Education			0.7
<HS	2 (4%)	1 (2%)
HS	13 (26%)	11 (23%)
Some college	23 (46%)	28 (57%)
≥College	12 (24%)	9 (18%)
Rurality			0.2
Least rural	15 (31%)	8 (16%)
Moderately rural	23 (47%)	24 (49%)
Most rural	11 (23%)	17 (35%)
Married or member of unmarried couple	14 (28%)	19 (39%)	0.6
Insured, of any kind	35 (70%)	31 (63%)	0.5
Smoking/Clinical History
Age began smoking	17.3 (4.3)	17.5 (4.2)	0.8
Baseline cigarettes per day (7 day avg)	18.4 (7.2)	15.1 (7.3)	0.03
Cigarette dependence (Heaviness of Smoking Index)	3.4 (1.3)	2.8 (1.5)	0.04
Lives with another smoker	20 (40%)	17 (35%)	0.3
Prior NRT use	34 (68%)	30 (61%)	0.5
Prior varenicline use	5 (10%)	4 (8%)	0.9
Concern about varenicline safety (1–5 range, 5 = most concerned)	2.1 (1.3)	2.1 (1.2)	0.9
History of anxiety	5 (10%)	4 (8%)	>0.9
History of depression	7 (14%)	9 (18%)	0.6
Prior E-cigarette use	24 (48%)	22 (45%)	0.8

Bold text refers to statistical significance at p < .05.

Varenicline Uptake

Restricting these outcomes to the varenicline group alone (only 1 person within control group reported varenicline use, on their own accord), any use of varenicline was 54% (n = 27; Week 2), 66% (n = 33; Week 4), 36% (n = 18; Week 8), and 24% (n = 12; Week 12). Rates of regular use, defined as using varenicline daily for each of the 7 preceding days were 38% (Week 2), 46% (Week 4), 20% (Week 8), and 10% (Week 12). Across the entire study, 76% and 58% of varenicline participants ever used, and ever used regularly. We were also interested in whether sampling of varenicline led to independent use (ie getting more, through their own healthcare provider/insurance), particularly in the latter weeks once their sample had been used. Among the intent-to-treat sample within the varenicline group, rates of independent use were: 8% (Week 2), 14% (Week 4), 12% (Week 8), and 14% (Week 12). Among users of varenicline samples at a given week, rates of independent use were: 15% (Week 2), 21% (Week 4), 33% (Week 8), and 58% (Week 12).

We were further interested in titration patterns, particularly since participants were given suggested but not required guidance on use. We sub-divided users of varenicline into 4 separate groups: (1) steady users or increasers (eg; 0.5 mg for a few days, then 1 mg and up to 2 mg for all 7 days), (2) increased then decreased (eg 0.5 mg, 1 mg, 2 mg, then back to 1 mg or 0.5 mg), (3) triers (eg a few days of use and then discontinued), and (4) sporadic (eg up, down, up, down). Of those who used varenicline at Week 2 (n = 27; 54% of varenicline group), the vast majority (67%) fell into steady users/increasers, and the remainder were spread across other use groups: (2) 15%, (3) 11%, and (4) 7%. The same held true at Week 4: of varenicline users (n = 33; 66% of varenicline group), the vast majority (58%) fell into steady users/increasers, and the remainder were spread across other use groups: (2) 6%, (3) 18%, and (4) 18%. We do not report titration patterns in latter weeks of follow-up given low overall rates of use (see above). At no time did anyone exceed the recommended maximum dosage (2 mg/day).

Safety

Focusing on the varenicline group alone, 26 participants (52%) reported a total of 90 adverse events, of which 35 (39%) were mild, 47 (52%) were moderate, and 8 (9%) were severe according to participant self-report. The most common adverse events were sleep disturbances (n = 28), nausea (17), agitation (9), depressed mood (8), and headache (5). In addition to these adverse events, follow-up assessment identified all instances of increases in depressive symptoms (increase of ≥10 points on PHQ-9 since baseline): there were 5 instances (5 people) within control group versus 12 instances (11 people) within varenicline group (3 instances with concurrent varenicline use). We also identified any endorsement of suicidal ideation as reported within the PHQ-9: there were 4 instances/people within control group versus 7 instances (5 people) within varenicline group (3 instances with concurrent varenicline use).

Hypothesized Process Measures of Sampling

Consistent with expectation, there was a significant time x group interaction for both motivation (p = .006) and confidence to quit (p = .02), with near parallel patterns of changes over time across both outcomes (which were generally equivalent at baseline); see Figure 2. Contrary to expectation, sampling varenicline did not differentially change attitudes towards varenicline itself, either in terms of perceived efficacy, perceived safety, or BAV (non-significant group x time interactions for all).

Figure 2.

Changes in motivation to quit (MTQ), confidence to quit (CTQ).

Cessation-Related Outcomes

Compared to the control group, participants in the varenicline group reported higher (numerically, but not statistically) rates for all outcomes related to quitting, including (1) incidence of any quit attempt throughout follow-up (60% vs. 49%; Adjusted Odds Ratio [AOR] = 1.47; 95% Confidence Interval [CI]: 0.64 to 3.42), (2) incidence of any 24 h quit attempt (52% vs. 35%; AOR = 2.24; 95% CI: 0.94 to 5.31), (3) point prevalence abstinence at 8 weeks (10% vs. 4%; AOR = 2.05; 95% CI: 0.35 to 12.16), (4) point prevalence abstinence at 12 weeks (18% vs. 6%; AOR = 2.59; 95% CI: 0.61 to 11.00), and (5) having ever reported 7-days quitting throughout the study; that is, “floating abstinence” (26% vs. 10%; AOR = 3.15; 95% CI: 0.97 to 10.22). In addition, there was a significant time x group interaction (p = .02), such that smokers in the varenicline reported greater reductions in CPD over time versus control (Figure 3). Participants in the varenicline group were significantly more likely to achieve 50% reduction in CPD, both at immediately following the sampling exercise (Week 4: 36% vs. 12%; AOR = 4.12; 95% CI: 1.39 to 12.17) and at the final follow-up (Week 12: 42% vs. 12%; AOR = 4.50; 95% CI: 1.56 to 13.01). Usage of the quitline was minimal in both groups: 8% versus 6% in varenicline versus control groups, respectively (AOR = 1.76; 95% CI: 0.34 to 9.04).

Figure 3.

Changes in smoking (CPD). *Cigarettes per day (CPD): average over seven days preceding each follow-up.

Sub-Group Analyses: The Role of Quit Motivation

We were interested in whether any of these outcomes differed as a function of baseline MTQ (ie whether benefits of sampling were restricted to only those smokers who expressed strong desire to quit), even if we were underpowered for analyses of treatment moderation. For all smoking-related outcomes, raw outcomes were higher among smokers who expressed higher versus lower MTQ. Nonetheless, the varenicline versus control effect sizes (AOR) were fairly similar across high versus low motivation groups. If anything, varenicline sampling had stronger effects among smokers with lower MTQ, for attempts to quit (AOR Low MTQ: 2.30 vs. AOR High MTQ: 1.11) and 24 h quit attempts (AOR Low MTQ: 2.76 vs. AOR High MTQ: 2.15), but stronger effects among smokers with higher MTQ for point prevalence abstinence at 12 weeks (AOR Low MTQ: 1.28 vs. AOR High MTQ: 5.51) and floating abstinence (AOR Low MTQ: 2.63 vs. AOR High MTQ: 3.30). Reductions in cigarettes per day (between treatment groups) were generally similar within subgroups of low versus highly motivated smokers (data not shown), consistent with patterns shown in Figure 3. Varenicline uptake was also generally comparable among smokers across motivation status.

Discussion

Prior to 2015, a sampling trial of varenicline would not have been possible, because of a FDA black box warning. The EAGLES trial²³ and a number of other persuasive reviews^44–46 have since demonstrated a favorable safety profile. The black box warning for varenicline has since been removed, which conceivably allows for greater population dissemination. This was our conceptual focus herein, testing whether an unstructured, user-driven experience of medication sampling could be applied to varenicline. While prior studies have examined novel approaches for varenicline delivery, we believe this is the first randomized trial to offer a naturalistic sampling context, through the mail, inclusive of smokers across the spectrum of MTQ, with no in-person interaction, circumscribed physician oversight, minimal instructions on use, and variable dosing as per participant self-determination.

Our findings clearly show that this is a feasible exercise. Medication uptake was moderate, with an appreciable number of participants using, using consistently, and moving on to independent use. These uptake outcomes are generally consistent with our prior trial of NRT sampling.¹² About 14% of smokers who received a sample of varenicline (58% of ever-triers) went on to get more of it, compared to 1 person (2%) in the control group. Titration patterns showed the modal trajectory of use conformed to guidelines, that is, slow increase up to 2 mg per day. Importantly, there were no reports of over-use. However, it is also worth noting here that a number of participants never tried it, prompting questions of cost/benefit that cannot be fully addressed here. More prescribed guidance on usage may yield stronger uptake. From a safety perspective, there were, as expected, higher rates of adverse events among smokers receiving varenicline, but the incidence and clinical severity of these adverse events were in line with prior trials, with no serious adverse events. However, it is also worth noting that there was a higher number of participants in the varenicline group who reported increases in depressive symptoms and/or any suicidal ideation. Importantly, each instance was addressed in real-time, and none were determined to be clinical emergencies. Whether these reflect medication-related adverse events or artifacts of withdrawal/changes in smoking (or both) is unclear.

Varenicline sampling clearly led to more favorable rates of all tobacco use outcomes: smoking reduction, quit attempts, and abstinence. Not all of these outcomes were statistically significant, and we caution against over-interpretation of results. As a pilot study, the focus is on yielding effect sizes to inform future large-scale clinical trials. From this metric (AOR; adjusted odds ratio), sampling yielded a 1.5- to 3-fold increase across all behavioral outcomes. It is equally compelling that these effects persisted to 3-month follow-up, likely well beyond the sampling experience. We had conceptualized the sampling experience as one that would promote motivation and confidence to quit (which was confirmed) and promote attitudinal shifts in favor of efficacy/safety (not confirmed).

Consistent with our earlier study of NRT sampling in primary care,¹² this trial demonstrates the potential benefits of medication sampling among smokers who express low MTQ. The provision of a concrete, immediately actionable tool for cessation is consistent with the notion of a cue to action within predominant models of behavioral change.⁴⁷ One simple but interesting study demonstrated that the mere mention of free cessation medication can shift the distribution of quit motivation,⁴⁸ offering a useful reminder that motivation is a fluid construct amenable to change through inducements to the quitting process.

Among study weaknesses, the limited sample size, generalizability of study sample, limited duration of follow-up and lack of biological verification of smoking stand out as most prominent. Remote/decentralized trials such as this are challenged by the logistics of remote capture of biological corroboration, yet the promise of mHealth strategies to address these challenges creates a compelling opportunity to do so in the future.⁴⁹ On the other hand, the generally favorable outcomes from this study point to the utility of offering brief samples of varenicline on a widescale basis, with circumscribed clinical oversight. Indeed a recent study did this within the quitline context.⁵⁰ It is also worth noting medication sampling as provided within this remote trial is fairly protected from larger contextual issues (COVID-19 as one example) that may further undermine access to in-person cessation services. Finally, from a regulatory perspective, our findings, if replicated in a larger trial, are supportive of alternative delivery modalities for varenicline.

In sum, this pilot randomized clinical trial demonstrated the feasibility and suggestive efficacy of varenicline sampling. A larger and more rigorous trial is warranted. Upon replication, varenicline sampling could serve as another strategy to engage more smokers in the cessation process sooner.

Supplementary Material

A Contributorship Form detailing each author's specific involvement with this content, as well as any supplementary data, are available online at https://academic.oup.com/ntr.

Funding

Supported in part by pilot research funding from the Hollings Cancer Center’s Cancer Center Support Grant (P30 CA138313) and with institutional support to MUSC via the National Center for Advancing Translational Sciences (UL1 TR001450). Drs. Dahne (K23 DA045766) and Smith (K01 DA047433) were supported by Career Development Awards from NIDA. Medication, but no other funding supported, provided from Pfizer.

Declaration of Interests

Drs. Carpenter, Gray, and Toll have received consulting honoraria from Pfizer. Dr. Carpenter has received consulting honoraria from Frutarom Inc, and Dr. Froeliger is a paid consultant for Promentis Pharmaceuticals, each for work unrelated to this research. Dr. Toll testifies on behalf of plaintiffs in litigation filed against the tobacco industry. Medication, but no other funding supported, provided from Pfizer.