Feasibility and acceptability of testing a menstrual-cycle timed smoking cessation intervention for women of reproductive age (Project Phase): Results of a pilot randomized control trial

Abstract

Introduction:

Menstrual phase influences cigarette smoking-related outcomes. Telephone-based cessation programs (e.g., quitlines) may incorporate the role of the menstrual cycle in an effort to tailor interventions for women.

Purpose:

The goal of this preliminary randomized clinical trial was to examine the feasibility and acceptability of timing quit date to menstrual phase in women in a quitline setting.

Methods:

We recruited treatment-seeking women with regular menstrual cycles between the ages of 18–40 years. Participants were randomized to the follicular phase (FP; quit date set 6–8 days post onset of menses) or standard of care (SC; no menstrual timing of quit date). All participants received four weeks of nicotine replacement therapy transdermal patch concurrent with six weeks of telephone-based counseling. We explored self-reported and biochemically-verified seven-day point prevalence abstinence at end-of-treatment and three-month follow-up.

Results:

Participants (n=119; FP: n=58, SC: n=61) were, on average, 33.4 years old and smoked 13.6 cigarettes/day. The median number of counseling sessions completed was 6 out of 6 available, and 66% of participants completed the intervention. Over 90% of participants reported they would recommend this study to friends/family. Cessation rates did not significantly vary by randomization.

Conclusions:

Results of this preliminary trial indicate that timing quit date to FP is an acceptable and feasible approach to address smoking cessation in women of reproductive age. While we observed similar smoking cessation rates between groups, this preliminary study was not fully powered to determine efficacy. Therefore, the feasibility and acceptability results indicate that a fully-powered efficacy trial is warranted.

INTRODUCTION

Although the prevalence of combustible cigarette smoking is lower in women than men, women are more susceptible to the harmful effects of smoking.^1–3 Women are also the primary source of tobacco smoke exposure in infants and children, increasing risk of acute infections (e.g., ear infections), chronic conditions (e.g., asthma), behavioral issues (e.g., attention deficit hyperactivity disorder), and death (e.g., sudden infant death syndrome) among offspring.^1,2 Further, even though men and women have similar levels of motivation for smoking cessation, women are less likely to quit smoking compared to men.^4,5 Despite decades of documentation showing less effectiveness of evidence-based smoking cessation interventions in women as compared to men,^3,5–7 interventions designed to the specific needs of women are lacking.

Telephone quitlines are an evidenced-based smoking cessation intervention that is a highly cost-efficient, scalable strategy to help individuals achieve smoking cessation.⁸ Standard quitline programs include telephone-based behavioral counseling sessions with trained counselors coupled with provision of nicotine replacement therapy (NRT). Considered a part of comprehensive tobacco control programs,⁹ quitlines are available in all states in the United States. Systematic reviews have shown that proactive telephone-based cessation services are an effective way to reduce cravings for and use of tobacco.^10–12 Nearly 60% of quitline callers are women,¹³ and studies show gender/sex differences in utilization of quitline services and cessation outcomes. For instance, our prior quitline research indicates that women, compared to men, complete fewer smoking cessation counseling sessions and are less likely to use NRT.¹⁴ Thus, identifying ways to improve compliance to quitline interventions in women of reproductive age will result in thousands more women quitting smoking every year.

One possible approach to increasing smoking cessation rates in women of reproductive age within a quitline setting may be to strategically set a quit date to a certain time during the menstrual cycle. Previous smoking cessation clinical trials on menstrual phase effects on cessation outcomes have been mixed, likely due to their inconsistent methodologies.^15,16 Specifically, the studies that observed luteal phase timed quit dates as favorable included one fully-powered randomized control trial with behavioral counseling¹⁷ and one secondary-data analysis that used bupropion as a cessation aid;¹⁸ neither used nicotine replacement therapy (NRT). In contrast, the two studies that observed follicular phase timed quit dates as favorable both used NRT as cessation aids, including one pilot randomized trial¹⁹ and one secondary-data analysis.²⁰ Finally, a secondary-data analysis that employed NRT and naltrexone observed no differences by menstrual cycle phase.²¹ To date, no studies have explored the role of the menstrual cycle within a quitline setting.

Here we hypothesize that setting a quit date during the follicular phase (FP; defined here as 6–8 days post onset of menses) will yield improved smoking cessation rates as compared to a standard of care (SC) approach within a quitline setting. We have made this hypothesis per four primary reasons. First, NRT efficacy may vary across the menstrual cycle due to greater nicotine absorption in the FP,²² modification of nicotine metabolism via estrogen,^23–25 and/or ovarian hormones modifying the neurobiological drug reward response to nicotine.^26,27 This may, at least in part, explain why NRT has limited effectiveness in women compared to men,^6,7,28–30 and, possibly, during FP.^{17–19,21,26} Second, given that most women of reproductive age relapse from a quit attempt within a matter of days regardless of which menstrual phase they attempt to quit in,¹⁷ reducing smoking-related symptomatology is also important. The FP has been linked to lower withdrawal symptoms, lower perceived stress, and decreasing premenstrual symptoms,^15,31 all of which have been identified as protective against relapse, regardless of menstrual phase.³² Third, progesterone, which peaks during the middle of the luteal phase has been shown to reduce the reinforcement value and favorable subjective effects of smoking/nicotine during an acute state of abstinence.^33–36 Thus, this higher progesterone level may translate into a reduced risk of a lapse resulting in a full relapse. Lastly and logistically, given FP is defined by the onset of menses, we may be able to easily operationalize and identify FP within a quitline setting, thus increasing scalability of our strategy.³⁷ To summarize, we hypothesized that setting a quit date during the FP within a quitline setting, as compared to SC, would be advantageous during an NRT-aided quit attempt due to: (a) increased absorption (and possibly efficacy) of NRT during FP, (b) diminished smoking-related symptomatology during FP, (c) the subsequent increasing progesterone during the early luteal phase (i.e., 1–2 weeks after a quit attempt has been made) that makes nicotine less reinforcing in the event of a lapse and perhaps reducing risk of a full relapse, and (d) the FP can be identifiable within a quitline setting, indicating it may be a feasible intervention.

As a first step to test our central hypothesis, the goal of this pilot randomized control smoking cessation trial is to explore the acceptability and feasibility of tailoring a woman’s quit date to her FP as an approach to improve smoking cessation compared to a SC condition (primary aim). We used a standard proactive telephone-based behavioral counseling approach paired with NRT. For our secondary aim, we examined preliminary efficacy of our menstrual-cycle timed intervention on self-reported and biochemically-verified smoking outcomes. Finally, given the final seven months of our two year study coincided with the emergence and spread of COVID-19, we descriptively explored the temporal patterns of our study outcomes in relation to the onset of the pandemic.

METHODS

Procedures.

The methodology used in this trial has been described in detail elsewhere.³⁸ In brief, we recruited women both reactively via national social media advertising and proactively through Arizona’s state quitline. During the last six months of our recruitment period, we expanded our social media advertising from the state of Arizona to the continental United States to ensure achievement of our enrollment goal. We targeted advertising to areas that had populations with more racial diversity and higher smoking prevalence in an effort to increase the diversity of our study sample. Eligible participants were women between the ages of 18 and 40 who self-reported smoking ≥ 5 cigarettes/day for ≥ 3 months, motivated to quit smoking, and regular menstrual cycle (i.e., length of 24–36 days). Exclusion criteria included recent (<3 months) pregnancy, breastfeeding, or use of exogenous hormones (including hormonal contraceptives). Eligibility was assessed by an online survey followed by a telephone interview. All enrolled participants provided informed consent and then completed baseline assessments via REDCap.³⁹

Those who met eligibility criteria were stratified by recruitment source (i.e., social media versus quitline referral) and then randomized (1:1 ratio) to the follicular phase group (FP; quit date occurred six to eight days post onset of menses) or the standard of care group (SC; no menstrual cycle timing consideration). Regardless of randomization, all participants completed a six-week intervention consisting of weekly telephone-based counseling sessions and were mailed a four-week supply NRT patch (two mailings were sent each containing two-week supply each and were dose tailored to their baseline smoking rate), with a quit date set in the third week of the intervention. Therefore, the FP participants started the six-week intervention one week prior to their expected onset menses, to allow the third week of the intervention (quit week) to occur six to eight days post onset of menses. To do this, study staff asked FP participants when they expected their next period and then participants were scheduled for their first session one week prior to the onset of menses. In contrast, the SC participants started the six-week intervention within two weeks of enrollment without regard to menstrual cycle timing.

Intervention staff included master’s degree-level health counselors. All counselors completed four weeks of training on the possible role of the menstrual cycle in smoking behavior change, evidence-based cessation counseling strategies, and motivational interviewing techniques. The counselors also met core competencies as outlined and assessed by the PIs. Counselors, who were not blind to randomization assignments, participated in weekly supervision with the PIs to review participant cases and problem solve participant issues as needed. Counseling sessions were recorded and reviewed to ensure ≥90% treatment fidelity. While the structure and frequency of the telephone sessions was the same between groups, participants in the FP condition were provided additional menstrual-based content (e.g., bloating) during each session.

Study assessments occurred at four weeks post quit date (end of treatment) and at three months post quit date, including self-reported seven-day point prevalence abstinence with biochemical confirmation via cotinine as measured in dried blood spots (DBS).^40,41 Specifically, participants self-collected DBS samples using mailed kits, and returned the samples to study staff via U.S. mail. DBS samples were then stored in a −20°C freezer. DBS samples were analyzed in bulk by the University of Minnesota (cotinine) and ZRT Labs (progesterone) using previously published approaches.^40,42 This project was registered on Clinicaltrials.gov (NCT03908320). All procedures were approved by the University of Arizona’s Human Subjects Protection Program.

Measures.

Baseline characteristics, including demographics, were collected at the baseline assessment. Our primary outcome was feasibility and acceptability. Feasibility outcomes included recruitment rate, retention rate, survey completion rate, and DBS samples received. In addition, for those randomized to the FP group, we assessed follicular phase based on cycle day (i.e., quit date occurred on menstrual cycle day 6–8) using progesterone value in DBS (i.e., < 2 ng/mL is consistent with follicular phase).³⁷ Acceptability was assessed by surveys at end of treatment and follow-up (see Table 2 for specific items). Following recent recommendations, smoking cessation (secondary outcome) was defined using the seven-day point-prevalence definition of abstinence with biochemical confirmation per cotinine value in DBS (<3 ng/ml indicative of abstinence).^40,43,44

Table 2.

Feasibility and Acceptability by Randomization

	Follicular Phase (FP) (n = 58)	Standard of Care (SC) (n = 60)1
Intervention Compliance
Counseling Sessions Completed, n (%)	Median=6.0	Median=6.0
- Session 1	56 (93)	42 (72)
- Session 2	54(90)	39 (67)
- Session 3	52 (87)	35 (60)
- Session 4	46 (77)	33 (57)
- Session 5	42 (70)	31 (53)
- Session 6	38 (63)	31 (53)
Average Duration of Sessions, mean minutes (SD)	34 (9.7)	30 (6.8)
NRT Patch Compliance	32 (55)	47 (78)
Data Collection Compliance
Time Points Completed, n (%)	Median =7.0	Median =7.0
- Baseline	58 (100)	60 (100)
- Session 1	44 (76)	54 (90)
- Session 2	36 (62)	55 (92)
- Session 3	36 (62)	52 (87)
- Session 4	34 (59)	46 (77)
- Session 5	34 (59)	43 (72)
- Session 6	35 (60)	42 (70)
- Week 4 (End of Treatment)	36 (62)	43 (72)
- Month 3 (Follow-up)	27 (47)	35 (58)
Dried Blood Spots Completed, n (%)	Median = 2.0	Median = 2.3
- Week −1	30(52)	41(68)
- Quit Day	31(53)	40(67)
- Week 1	25(43)	36(60)
- End of Treatment	21(36)	23(38)
- Follow-up	13(22)	16(27)
Progesterone Data on Quit Day	n=26	n=39
Progesterone Level (ng/mL), median (range)	0.2 (0.2, 7.0)	1.4 (0.2, 26.4)
Progesterone Level is < 2 ng/mL, n (%)	24 (92)	22 (56)
Menstrual Cycle Day of Quit	n=38	n=56
Median, interquartile range	7 (6,8)	16 (10.5, 21)
Quit day is 6 to 8 days post onset of menses, n (%)	29 (76)	3 (5)
Participant Satisfaction
Week 4 Assessment, n(%*)	n=36	n=42
- Study participation was “not at all” a burden.	32 (89)	36 (88)
- Completion of online survey was “not at all” a burden.	33 (92)	39 (93)
- Completing weekly dried blood spots was “not at all” a burden.	22 (61)	27 (64)
- Quit date was:   Sooner than I wanted it to be   Just at the right time   Later than I wanted it to be	11 (31) 20 (57) 4 (11)	7 (17) 28 (68) 6 (15)
- How willing would you be to complete this study again? (Definitely, probably yes).	34 (94)	40 (95)
Month 3 Assessment, n(%*)	n=27	n=35
- My coach discussed how the menstrual cycle may influence my craving levels and efforts in quitting. (Agree, strongly agree).	20 (80)	24 (75)
- The $100 compensation was fair.	24 (100)	28 (90)
- I would recommend this study to my friends and family who may be interested in quitting. (Somewhat agree, strongly agree).	25 (93)	32 (92)
- Overall, I am satisfied with the services I received from this study. (Somewhat agree, strongly agree).	22 (81)	32 (91)

^*Amongst those who responded.

¹One participant withdrew and asked that her data not be used.

Statistical Analyses.

We based our sample size on precision of feasibility outcomes.⁴⁵ A sample size of 112 yields precision (half-width of a 95% confidence interval [CI]) to estimate feasibility outcomes (e.g., recruitment and retention) to within ±9%, conservatively assuming a base rate of 50% (maximizing the standard error) and the formula for a 95% CI for a binary proportion. Baseline smoking characteristics, demographics, and feasibility outcomes were summarized with descriptive statistics. Comparison of cessation between the FP and SC groups at four weeks and three months post quit date were estimated with chi-squared tests both on the complete data and on a set where missing smoking data was imputed as returned to smoking.^40,43,44 We added an ad hoc analysis in which we duplicated the cessation analyses among a subgroup of those who were NRT adherent, given identified group differences in NRT adherence. Key feasibility outcomes were summarized with 95% CI. We also compared baseline characteristics of those who dropped out by three-months and those who did not.

RESULTS

Study Participants.

A total of 3,077 individuals expressed interest in the study. Of these, a total of 310 completed the telephone eligibility assessment (Figure 1). Of the 310 assessed, approximately 32% (n=99) were not eligible primarily due to menstrual cycle irregularity (n=48), use of exogenous hormones (n=22), and/or smoking < 5 cigarettes/day (n=16). The remaining 211 (68%) were sent informed consent and baseline surveys to be completed prior to randomization; 119 participants completed these items, were enrolled into the study, and were randomized with 58 assigned to FP and 61 assigned to SC. Participants were primarily non-Hispanic White (72.4% [FP] and 78.3% [SC]), were 33 years old, on average, and smoked an average of 13.5 (FP) and 13.8 (SC) cigarettes/day. Dropouts were, on average, significantly older, had more education, and smoked more cigarettes/day than participants who completed (Supplementary Table 1).

Figure 1. CONSORT Diagram.

LTFU = Lost-to follow-up

^a Reasons for exclusion include irregular menstrual cycles (n=52), use of exogenous hormones (n=22), smoking fewer than five cigarettes per day (n=16), contraindicated medical conditions (n=11), history of adverse reaction to transdermal nicotine (n=5), breastfeeding (n=3), current use of transdermal nicotine (n=2), unwilling to use transdermal nicotine(n=2), recent pregnancy (n=2), other (n=19).

^b Reasons for exclusion include irregular menstrual cycles (n=4), use of exogenous hormones (n=1), contraindicated medical conditions (n=12), other (n=5).

^c Reasons for exclusion include pre-existing medical conditions (n=12), irregular menstrual cycle (n=4), current use of NRT (n=3), already quit (n=1), use of St. John’s Wort (n=2), current quit attempt/unwilling to wait until quit day (n=2), hormonal birth control use (n=1), plans to become pregnant (n=1).

^d Reasons for exclusion included unwillingness to quit (n=1), exclusionary cycle length (n=6).

Feasibility.

Recruitment rate was 7.8 participants/month. Of those randomized (n=119), 99 (83%, 95%CI: 75%, 89%) started the intervention and 79 (66%, 95%CI: 57%, 75%) finished the six-week intervention. Retention rate through the three-month follow-up was 55% (95%CI: 46%, 65%). There were no significant differences in attrition by randomization (55.4% FP, 44.6% SC; p= 0.20). The median number of sessions attended by both arms was 6 out of 6 (Table 2). The average time for each counseling call was 34 (FP) and 30 (SC) minutes. NRT patch compliance was 55% (FP) and 78% (SC). The median number of completed data assessments was 7 out of 9 for both groups. The median number of completed DBS was 2.0 (FP) and 2.3 (SC) out of 5. Progesterone values on quit date ranged from 0.2 to 7.0 ng/mL (median 0.2 ng/mL) for the FP group and 0.2 to 26.4 ng/mL (median 1.4 ng/mL) for the SC group. Among the FP participants with complete data, 92% (24/26) had progesterone values consistent with FP (i.e., <2 ng/mL).

As displayed in Supplementary Figure 1, compliance rates varied over time. Specifically, compliance with DBS and follow-up surveys declined after the onset of the COVID-19 pandemic. In contrast, overall retention, NRT compliance, and counseling compliance stayed fairly stable.

Acceptability.

Of the 36 (FP) and 42 (SC) participants who completed satisfaction and acceptability surveys at Week 4, 88–89% agreed that study participation was “not at all” a burden, with 94–95% stating that they would probably or definitely be willing to participate again. Satisfaction at Month 3 was similar (Table 2).

Cessation.

Although our sample size was too small to estimate efficacy with confidence, we explored differences in cessation by randomization. Overall, no group differences in cessation outcomes were observed (Table 3). Biochemically verified seven-day point prevalence abstinence at four weeks post quit date was 10% (FP, n=10) and 17% (SC, n=12) (difference = −6.8, 95% CI: −28.8, 15.3). Similar results were observed based on self-reported data, imputing missing cessation values as returned to smoking, among NRT adherers, use of e-cigarettes, and at three months post quit date (Table 3). All 95% confidence intervals were wide and contained important differences favoring both the FP arm and the SC arm. Cessation rates initially increased during the onset of the COVID-19 pandemic, but then sharply declined during summer 2021. Though, comparison of pre-pandemic onset self-reported cessation rates (52% at end of treatment; 34% at follow-up) versus post-pandemic onset self-reported cessation rates (63% at end of treatment; 37% at follow-up) were similar.

Table 3.

Smoking cessation by randomization

	Follicular Phase (FP) n (%)	Standard of Care (SC) n (%)	Difference (95% CI)
End of treatment (4 weeks)
Biochemically verified	1/10 (10)	2/12 (17)	−6.7 (−34.8, 21.5)
Self-Report 7-day Point Prevalence (complete case)	19/36 (53)	25/42 (60)	−6.8 (−28.8, 15.3)
Amongst NRT adherers*	17/32 (53)	24/39 (62)	−8.4 (−31.5, 14.7)
Imputing missing as smoking	19/58 (33)	25/60 (42)	−8.9 (−26.3, 8.5)
Used e-cigarettes	6/36 (16.7)	1/42 (2.4)	14.2 (1.3, 27.3)
Follow-up (3 months)
Biochemically verified	4/5 (80)	2/7 (29)	51.4 (3.0, 99.9)
Self-Report 7-day Point Prevalence (complete case)	9/27 (33)	14/35 (40)	−6.7 (−30.7, 17.4)
Amongst NRT adherers*	7/24 (29)	14/34 (41)	−12.0 (−36.7, 12.6)
Imputing missing as smoking	9/58 (16)	14/60 (23)	−7.8 (−22.0, 6.4)
Used e-cigarettes	5/27 (18.5)	2/35 (5.7)	12.8 (−3.7, 29.4)

^*Defined as a second shipment of nicotine replacement therapy (NRT) patches was mailed to the study participant.

DISCUSSION

This preliminary randomized control trial aimed to examine the acceptability and feasibility of tailoring a telephone-based smoking cessation counseling intervention for women to include a target quit day timed to the FP as an approach to improve smoking cessation rates among women of reproductive age. Overall, participants found the intervention to be acceptable (e.g., >90% indicating they would recommend the intervention to friends/family). Further, based on both self-reported menstrual cycle day and progesterone values, the majority of participants randomized to FP had their quit date set in the follicular phase. This indicates that identifying the follicular phase via remote self-reported onset of menses within a quitline setting is feasible. Lastly, while we had a high-level of compliance to the behavioral counseling sessions (median of 6 sessions completed out of 6 provided) and a moderate-level compliance to NRT (58–78%), completion of surveys (median of 7 completed out of 9 requested), and overall retention in the study (55%). There was a relatively low-level of compliance for the self-collection of DBS (median of 2.0–2.3 collected out of 5 requested). However, many of these measures are similar or better than what we previously reported among female quitline callers, including use of cessation pharmacotherapy (71%), completion of five or more coaching calls (35%), and retention to follow-up assessment (conducted at month seven within the state quitline setting; 39%).¹⁴ Taken together, these observations indicate that tailoring a woman’s quit date to her FP is a feasible and acceptable approach to tailoring cessation interventions for women of reproductive age within a quitline-like setting. Additionally, efforts are needed to improve compliance to NRT and data collection procedures. This may include text message reminders, different participant payment structures tied to protocol adherence, reducing the number of time points, and/or re-training for DBS collection when a month or more occurs between collection time points.

Exploring group differences in smoking cessation outcomes (our secondary aim) the majority of our observed group differences were not statistically significant. We note that we were not powered to detect differences with confidence, and that the 95% confidence intervals were wide, containing effects that would favor either arm. Generally speaking, the percentage of abstinent participants was higher in the SC group (i.e., 17%−62%) compared to the FP group (i.e., 10%−53%). This may be related to a lower NRT compliance rate observed within the FP group versus the SC group (55% versus 78%, respectively). However, restricting the sample to NRT adherers revealed a similar pattern of higher cessation rates in the SC group compared to FC group. While both groups received similar recommendations for NRT patch use, it is possible our FP group counseling content (which included an emphasis on the role of ovarian hormones during the quit day and beyond) may have unintentionally minimized participants’ expected benefit of the NRT patch during the quit attempt, leading to less NRT use. Future efforts will explore ways to encourage the FP group to use NRT in addition to menstrual cycle timing to improve cessation outcomes. Additionally, the FP group also reported more e-cigarette use with 7% reporting use at baseline, and 16–19% reporting use during follow-up, compared to 0% use at baseline and 2–6% use during follow-up within the SC group. Overall, this observation suggests there may be interest in the use of e-cigarettes during cessation within women of reproductive age in a quitline-like setting, which is another area of future investigation. Next, it is important to note that our SC group is an active control condition that received an evidence-based smoking cessation intervention (specifically, proactive counseling combined with pharmacotherapy). Overall, our observed cessation rates in both groups were similarly effective to those observed in clinical trials, with 53–60% of participants self-reporting abstinence at the end of treatment (i.e., four weeks after assigned quit date) and 33–40% at three months after quit date. Further, though our biochemically-verified abstinence rates (e.g., 80% of the FP group at the three-month follow-up visit) were among a small portion of our full study sample (e.g., n=5 in FP group at the three month follow-up), these rates are similar to those observed in prior clinical trials of menstrual phase effects on NRT-aided cessation rates with biochemically-verified smoking abstinence rates. Specifically, in one trial, 69% of participants whose quit date coincided with the follicular phase were abstinent at nine weeks after quit date based on urine cotinine biochemical verification.²⁰ The second trial observed 32% of participants who were randomized to quit during the follicular phase were abstinent at two weeks post assigned quit date based on expired carbon monoxide biochemical verification.¹⁹ Further, our prior quitline research indicates that 39% of women self-report seven-day point prevalence abstinence seven months after enrolling in the Arizona state quitline.¹⁴ Here, our self-reported quit rates (a standard practice to assess quit outcomes in quitlines) of 33–40% at follow-up indicate that our outcomes are similar. Thus, overall, our intervention produced observed cessation rates are comparable to those observed in both clinical trials and quitlines.

Compliance to the behavioral counseling and average duration of time for phone sessions were comparable between the two groups in our study. This suggests that our intervention model, which educates women on the role of the menstrual cycle in smoking behavior, seamlessly integrates evidence-based coping strategies for smoking-specific and menstrual-cycle related triggers. Finally, while our intervention addressed biological (e.g., hormones) and psychological factors (e.g., negative affect), studies in this area can extend this framework to include impact of social factors (e.g., health literacy) on smoking outcomes. For instance, to use the knowledge of ovarian hormones to maximize favorable smoking behavior change, a minimal understanding of the menstrual cycle and the ability to use the information to make a quit attempt (i.e., health literacy) is necessary. Recent research indicates that health literacy plays an important role in reproductive knowledge and health behavior outcomes,⁴⁶ but its role in the ability to accurately report menstrual cycle related information (e.g., menstrual cycle length, date of last onset of menses) has not been examined. Accurate report of these data would be critical in the effective dissemination of this intervention.

This preliminary pilot study has provided us with several lessons. First, in relation to recruitment, a relatively small proportion of our sample was proactively recruited directly from the Arizona state quitline (≈8%). Our inclusion criteria limited the number of eligible quitline enrollees into our study. While almost 60% of callers to state quitlines are women, callers tend to be older (median age of 50), partly because quitlines only reach approximately 1% of tobacco users nationally.⁴⁷ Thus, our future work will benefit from increasing the number of collaborating quitlines to the recruitment plan – a strategy that has met with success in previous quitline-based studies.^48,49 Additionally, given the interest in this intervention as demonstrated by our successful recruitment of participants via social media, there appears to be additional opportunity for state quitlines to expand their reach, perhaps including target efforts for women of reproductive age. Further, for approximately the first 12 months of our 18 month recruitment period, we limited our target population to those who lived in Arizona. Given that Arizona has a relatively low prevalence of racial minorities, this resulted in a high prevalence of white participants in our study. During the last six months of our recruitment period, we successfully shifted our focus to metro areas that have higher prevalence of smoking and racial minorities (e.g., Philadelphia, Pennsylvania; Detroit, Michigan). This substantially increased the racial diversity of our sample, without any adverse effects on the logistics with distribution of study supplies. Lastly, in relation to the low compliance rate of the self-collection of DBS, a future efficacy trial may benefit from limiting number of self-collections to those time points that support the primary goal of the DBS (e.g., on quit date to determine menstrual phase and at end of treatment to biochemically verify smoking status). Additionally, it may be helpful to repeat the self-collection training after a period of time passes (e.g., a month). This may also improve self-collection compliance rates of DBS, which we have previously shown to be an acceptable and feasible approach to measure smoking status.⁴¹

Limitations are present. First, we observed evidence of selection bias as those who dropped out of our study differed from those who completed in terms of demographics (e.g., dropouts were older) and smoking behavior (e.g., dropouts smoked more cigarettes/day). Second, the majority of our sample reported they were non-Hispanic White and with a college education. Both limitations indicate limited generalizability. Third, we used self-reported onset of menses and progesterone values to identify follicular phase. It is possible that some of our participants may have had irregular menstrual cycles not identified by these measures. However, short of employing ultrasound to identify ovulation, which would not be feasible in this project or in a quitline setting, this is the best approach for identifying menstrual phase. Fourth, the end of our recruitment and data collection period occurred during the beginning of the COVID-19 pandemic, which impacted cessation and compliance rates. Lastly, since this was a feasibility and acceptability study, it was not powered to examine efficacy of the intervention. Therefore, a full-powered efficacy trial is needed to identify group differences.

In conclusion, the results of this preliminary randomized control trial indicate that women of reproductive age are willing and able to have their quit date set during the follicular menstrual phase. Future analyses will include the role of potential effect modifiers (e.g. anxiety sensitivity, urge coping) which will further inform the development of a potential quitline intervention. A future, fully powered efficacy trial is warranted, especially given this type of tailoring within a quitline setting comes at little to no cost.

Supplementary Material

1

Supplementary Table 1. Comparison of participants who dropped out versus those with 3 month follow-up data

Supplementary Figure 1. Compliance, Retention, and Cessation Rates over Time

Table 1.

Demographic and Baseline Characteristics by Randomization¹

	Follicular Phase (FP) (n = 58)	Standard of Care (SC) (n = 60)2
Demographics
Age (years), mean (standard deviation [SD])	33.5 (5.2)	33.4 (5.2)
Race/Ethnicity
Non-Hispanic White	42 (72.4)	47 (78.3)
Non-Hispanic Black	6 (10.3)	6 (10.0)
Hispanic	4 (6.9)	3 (5.0)
American Indian	3 (5.2)	1 (1.7)
Multi-racial	3 (5.2)	3 (5.0)
Education
High school or less	12 (20.7)	14 (23.3)
Some college	33 (56.9)	23 (38.3)
College graduate or more	13 (22.4)	23 (38.3)
Smoking-Related Characteristics
Cigarettes/day, mean (SD)	13.5 (4.9)	13.8 (6.0)
Smoke within 30 minutes of waking	43 (75.4)	47 (81.0)
Number of previous quit attempts, mean (SD)	0.9 (1.2)	1.2 (1.8)
Uses e-cigarettes daily or weekly	4 (6.9)	0 (0)

¹Values shown are n (%) unless otherwise noted.

²One participant withdrew and asked that her data not be used.

Highlights.

• Menstrual cycle phase may influence cessation outcomes.

• The role of the menstrual cycle on cessation in a quitline setting is unknown.

• We demonstrates feasibility and acceptability of this approach.

• Future work needs to determine the efficacy of novel quitline intervention.