Increases in Endogenous Progesterone Attenuate Smoking in a Cohort of Non-Treatment Seeking Women: an exploratory prospective study.

Abstract

Despite advances in prevention and treatment, cigarette smoking remains a leading cause of preventable death in the United States. While men and women are equally likely to attempt to quit smoking cigarettes, women are far less likely to achieve abstinence both during and following cessation treatment. Recent evidence suggests that ovarian hormone levels may play a role in successful abstinence attempts in women smokers. The primary goal of this exploratory prospective observational study was to estimate the association between within participant levels of progesterone and estradiol with associated cigarettes smoked per day in adult women smokers (n=115). The primary study outcome was self-reported cigarettes smoked per day (CPD) during a two-week observational period collected using a daily smoking diary. Additionally, participants collected saliva daily, from which hormone levels (progesterone and estradiol) were derived.

Higher within-participant progesterone levels were associated with a significant decrease in CPD (p=.008) whereas within-participant estradiol levels were unrelated to CPD (p=.25). Regression models indicated a single change in the trajectory of smoking behavior for both within participant progesterone and estradiol. When progesterone values were below the change point, there was a significant inverse relationship between within-participant progesterone levels and smoking behavior (p=.025) while the relationship was attenuated for higher within-participant progesterone levels (p=.59). The effect of estradiol on smoking behavior was not significant when it was either below (p=.92) or above (p=.16) the change point. Higher within-participant levels of progesterone but not estradiol, are associated with reduced CPD in non-treatment seeking women smokers.

1. Introduction

In spite of advances in prevention and treatment, cigarette smoking remains a leading and substantial preventable cause of death in the United States (US) and globally^1,2. While prevalence rates have been dropping steadily, 34 million US adults remain smokers and nearly half a million die annually from tobacco-related illnesses^1,3. Although pharmacological and behavioral interventions are effective, relapse to smoking following cessation treatment occurs in up to 80% of patients within the first year⁴. Men are more likely than women to be current cigarette smokers⁵. Although women are equally likely to make a quit attempt, they are far less likely than men to achieve abstinence both during and following cessation treatment^6,7,8. Women have had lower reported rates of cessation in community-based trials as well as in bupropion and nicotine replacement therapy treatment efficacy trials^9,10,11,12. Thus, most of the existing evidence suggests that women are less likely to maintain abstinence than men.

There is considerable evidence that women’s subjective responses to drugs of abuse, including nicotine, are affected by ovarian hormones (progesterone and estradiol)^13,14,15,16. In a small laboratory study of women smokers, participants received either oral progesterone or matching placebo while they were in the early follicular phase of their menstrual cycle; administration of the progesterone at this time served to simulate the levels that would typically be reached during their luteal phase¹⁷. Following two puffs on a cigarette, women given progesterone vs. placebo reported attenuated craving. In a follow up laboratory study with a similar design, progesterone attenuated self-reported “drug-liking” and enhanced self-reported “bad-effects” of nicotine in both women and men smokers¹⁸. Additionally, increased doses of progesterone reduced tonic (non-cue elicited) craving but did not reduce ad-libitum smoking as compared to placebo¹⁹.

Allen and colleagues examined menstrual phase effects on smoking behavior and relapse in women attempting to quit smoking²⁰. A greater latency to relapse was observed in women who quit during the luteal vs. follicular menstrual cycle phase. A study from our group examined the association between naturally occurring fluctuations in progesterone and estradiol and laboratory-based ad-libitum smoking behavior in women²¹. We observed that decreases in progesterone (P) and estradiol (E) over the 10-day period prior to the laboratory session were associated with increased puff intensity (a proxy for smoking motivation). Additionally, decrease in the ratio of progesterone to estradiol was associated with greater number of puffs and weight of cigarettes smoked. Additionally, ovarian hormones have also been shown to impact cessation from smoking. In a recent smoking cessation study, our research team demonstrated that women with increasing in progesterone levels (i.e. early luteal phase) had a significantly greater likelihood to report 7-day abstinence from cigarettes as compared to women with either no change or decreasing progesterone levels²². Pang and colleagues collected repeated salivary progesterone and estradiol at three laboratory sessions occurring during distinct ovarian cycle phases in abstinent women²³. Within-participant increases in progesterone were associated with decreased negative affect and urges to smoke. These results taken together may suggest that both magnitude and direction of within-individual progesterone changes may significantly affect smoking behavior.

The primary study aim of this study was to assess the relationships between daily ovarian hormone levels (progesterone and estradiol) and smoking behavior in non-nicotine deprived and non-treatment seeking women smokers. Specifically, we hypothesized that higher subject specific endogenous progesterone levels will be associated with a decrease daily smoking.

2. Materials and Methods

2.1. Participants

Non-treatment seeking, cigarette smoking women (n=115; ages 18–45) were recruited from the community from March 2013 through May 2017. Study eligibility was based on the following inclusion criteria: 1) smoke an average of 5 cigarettes/day for at least the past 6 months, 2) submit a breath carbon monoxide (CO) sample at screening of at least 5 parts per million (ppm), 3) be post menarche and pre-menopausal with regular menstrual cycles between 25 and 35 days, and 4) if recently pregnant, participants had to be at least 3 months post-delivery/breast feeding. Exclusion criteria included: 1) any serious or unstable medical (e.g., cancers, or psychiatric condition), 2) meeting criteria for post-traumatic stress disorder, 3) any medication (e.g., propranolol) that may interfere with psychophysiological (e.g., heart rate) monitoring during the laboratory session, 4) current substance dependence other than nicotine (or caffeine withdrawal) in the past month (as measured by the Structured Clinical Interview for Diagnostic and Statistical Manual-IV Axis I Disorders, 5) use of other tobacco products, and 6) participants who were pregnant, breast feeding, post hysterectomy or bilateral oophorectomy, or taking birth control or hormone replacement medication that would affect the menstrual cycle²⁴. Participants could enter the study at any time during their menstrual cycle. Study procedures were reviewed and approved by the Institutional Review Board at the Medical University of South Carolina (MUSC).

2.2. Procedures

Following informed consent and determination of eligibility, participants began a 14-day observational study. Participants collected daily salivary hormone samples (progesterone and estradiol) every morning, completed a daily morning cigarette report on a mobile app, and completed semi-random Cue Reactivity Ecological Momentary Assessment (CREMA; smart phone app for collecting participant responses/outcome measures in the field) sessions on the app²⁵. Related details and results from CREMA have been published elsewhere^26,27. Participants received monetary compensation for completing the baseline assessment ($100), attending the day 7 and day 15 visits ($50) and compliance with return of salivary samples for hormone measurement ($5/per) as well as completion of morning smoking reports ($1.25/per). This study was a part of a larger two stage study that examined a) the relationship between ovarian hormones and CREMA with daily smoking and b) the effects of oxytocin on response to a laboratory stressor. The laboratory session was scheduled for the final day of the second week of participation (Day 15) where stress reactivity smoking resistance in the presence of a treatment (oxytocin vs. placebo) was completed²⁸.

2.3. Measures

Screening Assessments. Standard demographic, medical and psychiatric histories were gathered at the screening visit, in addition to current tobacco use, smoking patterns, and previous smoking quit attempts and history. Nicotine dependence was assessed via the Fagerström Test for Nicotine Dependence²⁹. Participants completed the Menstrual History Diary to assess the timing of the menstrual cycle for the 90-days prior to study entry and to track their cycle during study participation. Retrospective Smoking Measures. Cigarettes per day (CPD) was assessed via the Timeline Follow-Back (TLFB) for the 30 days leading up to the two-week monitoring period³⁰. TLFB data were also collected at the final visit (~Day 15) to cover all days in the study. Daily Morning Report. Participants were prompted on the mobile app each morning at a time of their choosing to report the number of cigarettes smoked and potential menses onset (yes/no) during the previous day. Ovarian Hormone Collection and Assays. Participants collected saliva samples 30 minutes after awakening. Samples were placed into containers supplied by study staff and stored in participants’ home freezers until they were delivered to study staff at a subsequent visit. The samples were assayed at the South Carolina Clinical & Translational Research (SCTR) Institute laboratory (MUSC) using competitive enzyme immunoassay kits manufactured by Salimetrics™. Importantly, data supplied by the manufacturer indicates that the saliva measures of progesterone and estradiol are similar with respect to their (1) coefficients of variation (coefficients for high and low levels of progesterone were 4.8 and 9.0, respectively, whereas coefficients for high and low levels of estradiol were 7.0 and 8.1, respectively), and (2) correlation with serum assay levels (r’s=.8, p’s ≤ .001 for saliva/serum progesterone and saliva/serum estradiol). Addtionally, the data indicates that the concentration of progesterone and estradiol that can be distinguished from 0 is at least 5.0 pg/mL and 0.1 pg/mL, respectively. These daily home collection and storage procedures have been used in numerous previous studies with excellent participant adherence and sample quality^31,32.

2.4. Study Outcome

The primary study outcome was self-reported cigarettes smoked per day during the two-week observational phase of the study collected using the mobile app.

2.5. Statistical Methods

2.5.1. Overview

Hormone levels were centered at each participant’s mean³³ (Person Mean Centered; PMC, where higher levels indicate that hormones are greater than the participant’s average). Prior research suggests, the relationship between PMC hormone levels and smoking behavior may vary during the menstrual cycle leading to a non-linear relationship and variation in the correlation between the two^21,23.

Statistical analysis were performed in three steps. First, the relationship between hormones and self-reported CPD were assessed assuming a linear relationship using a generalized linear mixed effects model. Second, non-linear relationships between CPD and hormone levels were examined through locally estimated scatterplot smoothing (LOESS) utilizing predicted values from the generalized linear model. In the third step, as suggested by the LOESS fit from the previous step, a piecewise linear regression model was fitted and allowed for differing levels of PMC hormones to have varying associations with CPD via a model derived change point. All statistical analyses were conducted using SAS version 9.4 (SAS Institute Inc. Cary, NC, USA). Significance was set at a 2-sided α≤0.05.

2.5.2. Baseline Analysis and Participant Characteristic.

Baseline demographics, clinical and tobacco use characteristics were tabulated for study participants. Additionally, baseline, day level and session level characteristics were independently assessed for association with CPD. Variables that indicate association with outcomes were retained for model development (P<.05). These covariates in addition to design covariates (study day, menstrual cycle day) were included in all study models. Additionally, overall participant average hormone levels were included in the models to adjust for phenotypic variability across participants.

2.5.3. Linear Relationship between Ovarian Hormones and CPD

To assess the linear association between daily PMC progesterone and estradiol levels and concurrent CPD, generalized linear mixed effects models (GLMM) including a random intercept term were fit. Within participant correlation structures were modeled using multiple structures and final models were chosen while minimizing Akaike information criterion (AIC). Further, analysis of the effects of grouped hormone levels (high/low) and CPD were similarly assessed using GLMMs.

2.5.4. Non-Linear Modeling of Ovarian Hormones and CPD

Locally estimated scatterplot smoothing (LOESS) methods were implemented to determine if the relationship between PMC hormones and CPD is different at varying levels of each hormone. When data indicated that there may be a piecewise relationship between the variables, generalized non-linear mixed effects models were used to estimate the location of the change in association (change point location). Resulting change point location(s) were included in piecewise regression models allowing for correlated measures within each participant as well as random intercept terms while estimating distinct slopes and standard errors for each linear segment.

Examination of the piecewise regression models indicated varying effects of PMC hormones on either side of the change points. In addition to the regression analysis, these change points were used to combine hormones into participant specific groupings (High/Low) to assess the additive effect of high within participant levels of both progesterone and estradiol on smoking behavior as compared to low within participant levels or combinations of either.

All continuous covariates were mean centered prior to model inclusion (age, BL CPD etc.). All models included baseline CPD, study day, menstrual cycle day (centered at most recent menses) and PMC hormone levels (progesterone and estradiol). Further, models included variables significantly correlated with smoking behavior during the study (participant age and concurrent alcohol consumption) as well as the overall mean participant hormone level as a surrogate phenotype measure for generally high or generally low hormone levels relative to other participants. Additionally, uncentered hormone values were stratified into quartiles and GLMMs were fit to assess the overall relationship between hormone levels and reported CPD.

3. Results

3.1. Study Participants

A total of 115 women were included in the study and 104 (90.4%) provided viable saliva samples for hormone measures (Table 1). Participants were on average 31 years old (SD=7.1) and primarily Caucasian (62.5%). At study entry, participants reported moderate nicotine dependence (FTND=4.6, SD=2.0, range 0–8) and smoked an average of 15 cigarettes per day (30-day TLFB; range 5–40 CPD). Of the 104 participants included in this study, there were a total of 1387 (95% compliance) viable salivary hormone samples collected (1387 progesterone and 1385 estradiol samples) while a total of 1142 morning reports, including self-reported CPD for the previous day, were collected (77% overall compliance; median report=13.7 CPD; Range 1–43 CPD during the study). There were 1136 corresponding data points with both hormone and morning report smoking data recorded (participant median=12 reports, Inner quartile range=9–14 reports). Seventy participants (67.3%) noted any alcohol consumption during the 30 days prior to screening with an average of 5.4 drinking days in those that drank (SD=5.6; Drinks per drinking day: 3.7 (2.6)). Further, 39 participants (37.5%) noted the use of illicit substances at study screening (average of 5.5 (SD=8.0) use days). Of those 39, 36 (92%) noted using cannabis exclusively while 3 additional participants each noted only 1 cocaine use day during the prior 30 days. No data were missing on other study covariates.

Table 1.

Baseline Demographics.

	Study Participants (n=104)
Age
M	31.3 (SD=7.1)
Race (%)
White/Caucasian	62.5
Black/African American	34.6
Biracial	1.9
Other	1.0
Ethnicity (%)
Hispanic or Latino	4.8
Not Hispanic or Latino	95.2
Marital Status (%)
Single, Divorced, Separated, Widowed	81.7
Married	18.3
Highest Level of Education (%)
No High School Diploma or GED	13.5
High School Diploma or GED	26.9
Vocational School Graduate	1.0
Some College	32.7
Associate’s Degree or Equivalent	10.6
Bachelor’s Degree or Equivalent	12.4
Graduate or Professional Degree	2.9
Employment/Income
Currently unemployed (%)	28.9
Income less than $25,001 (%)a	58.3
Cigarette Use (Prior to study enrollment)
Average Cigarettes/Day (30-Day TLFB)	14.9 (SD=6.8)
Nicotine Dependence (FTND)	4.7 (SD=2.1)
Other Substance Use (Prior to study enrollment)
Any Alcohol Use % (n)	67.3% (70)
Days of Alcohol Useb	5.4 (5.6)
Drinks per drinking dayb	3.7 (2.6)
Other Drug Use % (n)	37.5% (39)
Days of Other Drug Useb	5.5 (8.0)

Data are shown as means and standard deviation for continuous variables and percentages for categorical variables. Note.

^aIncome data not available for 2 participants.

^bOf those reporting any use.

3.2. Correlates with Daily Smoking Behavior

Bivariate analysis of time variant and invariant characteristics was performed to assess individual associations with daily smoking behavior. Greater baseline smoking (CPD: β=0.71; SE=0.06; t₉₇=12.3; p<.001), participant age (β=0.21; SE=0.08; t₁₀₃=2.5; p=.014) and any concurrent alcohol consumption (during the same 24 hour period; β=1.16; SE=0.37; t₁₁₀₁=3.1; p=.002) were associated with increased CPD.

3.3. Within-Participant Smoking Behavior

In a linear model assessing the smoking level response to within-participant increases in hormone levels, within-participant increases in progesterone levels (Table 2; PMC; 1 SD increase=86 ng/ml) were associated with a significant decrease in CPD (β=−0.34; SE=0.13; t₉₉₁=−2.7; p=.008) while within-participant increases in estradiol levels (PMC; 1 SD increase=0.77 ng/ml increase above the participant mean value) were not associated with a significant change in CPD (β=0.14; SE=0.12; t₉₇₉=1.2; p=.25).

Table 2.

Summary of the linear and piecewise linear models for the relationship between PMC Hormones and CPD.

		Piecewise Regression
Factor	Linear Regression	PMC Progesterone ≤30 ng/ml	PMC Progesterone >30 ng/ml	PMC Estradiol ≤0.77 ng/ml	PMC Estradiol >0.77 ng/ml
Progesterone - Person Mean Centered*	−0.34 (0.13)*	−0.45 (0.20)*	−0.16 (0.32)	−0.35 (0.13)*	0.14 (0.52)
Progesterone - Participant Mean*	−0.45 (0.93)	−0.47 (0.94)	−0.18 (1.00)	−0.43 (0.93)	0.06 (1.06)
Estradiol - Person Mean Centered*	0.14 (0.12)	0.03 (0.19)	0.32 (0.48)	0.03 (0.18)	0.51 (0.39)
Estradiol - Participant Mean*	1.25 (1.24)	1.18 (1.24)	1.47 (1.34)	1.15 (1.24)	1.64 (1.31)
Cigarettes per day (BL)	0.73 (0.06)*	0.71 (0.06)*	1.00 (0.45)*	0.71 (0.06)*	1.20 (0.51)*
Age (1 yrs)	0.03 (0.06)	0.03 (0.06)	0.32 (0.45)	0.03 (0.06)	0.52 (0.50)
Any Concurrent Alcohol Use (yes vs. no)	1.41 (0.37)*	1.37 (0.38)*	1.66 (0.59)*	1.37 (0.37)*	1.86 (0.62)*
Cycle Day	−0.01 (0.02)	0.00 (0.02)	0.17 (0.44)	0.00 (0.02)	0.50 (0.50)

Data are shown as regression coefficient and associated standard error. Progesterone and estradiol measures are noted for a 1 SD unit change; PMC Progesterone, 1 SD=86 ng/ml; Participant mean progesterone, 1 SD=161 ng/ml; PMC estradiol, 1SD=0.77 ng/ml; Participant mean estradiol, 1 SD=2.21 ng/ml.

^*P < .05

Locally estimated scatterplot smoothed (LOESS) regression models indicates a single change point in the association of smoking behavior between 0 and 100 ng/ml above the PMC progesterone measurement (Figure 1). A non-linear mixed effects model was used to estimate the change point location. A single change point location at 30 ng/ml of PMC progesterone was noted in the model and piecewise regression functions were fit. In the adjusted model, when PMC progesterone values were below the change point value (low relative to participant mean), there was a significant inverse relationship between within-participant progesterone levels and smoking behavior (β=−0.45; SE=0.20; t₉₈₁=−2.2; p=.025) while the linear relationship was attenuated for higher participant specific progesterone levels (β=−0.16; SE=0.32; t₉₇₉=−0.5; p=.59). This indicates that higher participant specific levels of progesterone had a significant effect on smoking behavior. However, beyond the change point (PMC~30 ng/ml), the level of progesterone remains associated with lowered CPD but increases in progesterone no longer have an additive influence on behavior. Similar nonlinear and piecewise regression analyses of estradiol values revealed a possible change point location at approximately PMC~0.5 ng/ml (Figure 2). In the adjusted models, when PMC estradiol was below the change point, the effects of estradiol on smoking behavior were not significant (β=0.03; SE=0.18; t₉₉₃=0.1; p=.92) while the relationship was insignificant, yet increasing for higher within participant estradiol levels (β=0.51; SE=0.39; t₉₉₃=1.4; p=.16).

Figure 1:

Within participant relationship between smoking behavior and changes in participant progesterone.

Data shown as model based predicted means and 95% confidence interval from locally estimated scatterplot smoothing (LOESS) regression models.

Figure 2:

Within participant relationship between smoking behavior and changes in participant estradiol.

Data shown as model based predicted means and 95% confidence interval from locally estimated scatterplot smoothing (LOESS) regression models.

To further assess the combined effect of within-participant progesterone and estradiol on smoking behavior, we grouped observations according to whether hormones were above or below the PMC change point values (4 groups; 1. Low P/Low E, 2. Low P/High E, 3. High P/Low E, 4. High P/High E). Although the overall effect of grouping did not reach statistical significance (F_3,992=2.5, p=.059), there was a marked difference between observations with low within-participant progesterone as compared to the high within-participant progesterone, specifically when estradiol was lower (p<.05; Figure 3).

Figure 3:

Relationship between combined high and low participant levels of progesterone and estradiol and smoking behavior.

Data shown as adjusted model based mean CPD and associated Standard Errors

* P<.05 as compared to High P / Low E group (p=.015, p=.028).

3.4. Between-Participant Smoking Behavior

The mean progesterone level during study treatment was 162.2 ng/ml (SD=117.9; range=1.63–758.1) and the mean estradiol level during the study was 2.18 ng/ml (SD=1.10; range=0.36–9.58). Study reported cigarettes per day are shown stratified by quartile of un-centered progesterone and estradiol values in table 3. There was no significant association between reported CPD across quartiles of progesterone (F_3,1011=1.1; p=0.35) or estradiol (F_3,1007=1.4; p=0.23). Similarly, in adjusted GLMMs, participant mean progesterone (F_1,96.5=0.2; p=0.63) and estradiol levels (F_1,104=1.0; p=0.31) were not significantly associated with CPD (Table 2).

Table 3.

Summary of the relationship of CPD and uncentered progesterone and estradiol.

Progesterone		Estradiol
Quartile	CPD	Quartile	CPD
1st Quartile (<73.2 ng/ml)	14.1 (0.5)	1st Quartile (<1.39 ng/ml)	13.5 (0.5)
2nd Quartile (73.2–131.66 ng/ml)	13.9 (0.5)	2nd Quartile (1.39–2.00 ng/ml)	14.0 (0.5)
3rd Quartile (131.66–222.92 ng/ml)	14.1 (0.5)	3rd Quartile (2.00–2.66 ng/ml)	14.3 (0.5)
4th Quartile (>222.92 ng/ml)	13.4 (0.5)	4th Quartile (>2.66 ng/ml)	13.8 (0.5)

CPD data are shown as adjusted regression coefficient and associated standard error (SE).

4. Discussion

This is the first study to report on the association between daily endogenous ovarian hormone levels and the smoking behavior of untreated women. Higher levels of participant specific progesterone were associated with decreased CPD. The lowest reported CPD occurred when within-participant progesterone was high and within-participant estradiol was low. Furthermore, progesterone-related decrements in CPD occurred when progesterone increased from low vs. relatively high within-participant levels. This indicates a possible inflection point in the progesterone curve when the rate of hormone increase begins to slow down; likely during increases in the luteal phase. These findings converge with our previous work showing that progesterone levels in combination with the magnitude and direction of change of progesterone may play an important role in smoking behavior²³. Specifically, more rapid early luteal increases in progesterone appear to yield the greatest dampening effect on smoking and that this dampening effect diminishes as progesterone reaches asymptotic levels later in the luteal phase.

Our results are consistent with a growing body of preclinical and clinical research that indicates progesterone may attenuate drug reward and drug-seeking behavior^18,19,20,34. While sparse, the few nicotine studies are in agreement with the studies involving non-nicotine reinforcement^35,36. Recent work by Tosun et al. involving exogenous progesterone for smoking abstinence noted a greater than 2-fold increasing odds of abstinence as well as a delay in relapse in women receiving 200 mg of progesterone twice daily as compared to placebo³⁷. Pang et al. noted that higher within participant progesterone levels were associated with decreases in negative affect and smoking urges²⁴. We hypothesized that this may be a pathway between changes in progesterone and smoking behavior. Our work shows that while women are more responsive to stressful cues than men, cue-elicited negative affect and stress are not associated with smoking in the natural environment³⁸.

Limitations:

One limitation of the present study was that the identified relationships/effects were relatively small. However, these effects seem especially noteworthy as they were observed in non-treatment seeking daily smokers who were not asked to change their smoking behavior and who did not receive any treatment. Further, without complete menstrual cycle data for each participant, within-participant comparisons between times of rapid increases vs. decreases in ovarian hormones were not possible. Although the association between progesterone and smoking behavior was significant, similar associations with estradiol were not. Given that the estradiol assay is more sensitive than the progesterone measure (see assay comparative metrics and precision data in section 2.3 Measures above), it would seem that if associations between progesterone levels and smoking are detected, similar associations between estradiol levels and smoking should be detected if they exist; however, we did not. Additionally, although each participant collected salivary hormone samples daily, samples were delivered to study staff on a weekly basis, leaving the opportunity to have reduced data reliability; specifically, the opportunity to complete all hormone samples on a single day. Our study staff worked diligently to identify these instances and flag them in the data file. Statisticians reviewed each instance and it was determined that all data should be included in the analyses. Further, the hypothesis tested in this analysis was not pre-registered or specified prior to undertaking the work and results should be considered exploratory.

A greater understanding of the role of fluctuations in progesterone and estradiol in smoking reward and relapse may aid in the development of sex-specific interventions to enhance treatment response in women smokers. One way this might be achieved is by strategically timing and dosing exogenous hormone administration such that it simulates the changes that occur during the first half of the luteal phase, thereby augmenting the treatment yield of existing first-line cessation interventions among women. Moreover, since we collected an average of 14 contiguous days of hormone data for each participant, we did not capture the full menstrual cycle hormone profile of any participant. However, these results combined with prior findings indicate that making a quit attempt during the early to mid-luteal phase, when the rate of progesterone increase is approaching its maximum, may increase the probability of abstinence. Additionally, the present findings suggest that administering a static dosing regimen of exogenous progesterone for smoking cessation may be suboptimal for women smokers. Rather, these findings point to the possibility that a strategically timed quit date together with a carefully timed escalating dosing regimen of exogenous progesterone might serve to increase the likelihood of successful abstinence. Alternatively, the strategic timing of a quit date (luteal menstrual phase) might also serve to improve the efficacy of conventional first line cessation interventions.