Tobacco Cigarette Smokers Who Endorse Greater Intolerance for Nicotine Withdrawal Also Report More Severe Insomnia Symptoms

Abstract

It has been suggested that nighttime nicotine withdrawal may help to explain why tobacco cigarette smokers are more likely than nonsmokers to experience clinically significant insomnia. There is also reason to believe that intolerance for withdrawal symptoms could play a role in withdrawal-related sleep disturbance. However, we are not aware of any previous research that examined whether smokers who endorse greater intolerance for smoking abstinence also report greater difficulty initiating and/or maintaining sleep. To address this question, 224 adult cigarette smokers (42.9% female, M_{cigarettes per day} = 21.3) completed the baseline portion of an experimental study that included assessment of current/historical smoking behavior, perceived intolerance for smoking abstinence, and insomnia severity and impact on functioning. Results indicated that, after accounting for general distress intolerance and sociodemographic factors, smokers who endorsed greater intolerance for nicotine withdrawal also reported greater insomnia severity and impact. Logistic regression further revealed that, for every one-point increase in nicotine withdrawal intolerance scores, smokers were nearly twice as likely to score above threshold for clinically significant insomnia (p = .001). Collectively, these initial findings suggest that intolerance for nicotine withdrawal may warrant consideration as a potentially modifiable mechanistic factor in comorbid insomnia and nicotine/tobacco dependence.

Approximately one-third of American adults report insomnia symptoms (i.e., difficulty with initiating or maintaining sleep and/or waking up too early; Olfson et al., 2018), and an estimated 10–20% meet Diagnostic and Statistical Manual criteria for insomnia disorder, (i.e., the sleep disturbance causes clinically significant distress and/or impairment of daytime function; American Psychiatric Association, 2000, 2013; Chung et al., 2015; Roth et al., 2011), depending on the edition used. Insomnia symptoms present a tremendous public health burden, contributing to chronic illness (e.g., cardiovascular disease, chronic pain, depression; Herrero Babiloni et al., 2019), substance use (e.g., Taylor et al., 2003), and lost quality of life (e.g., Olfson et al., 2018). Untreated insomnia costs over $1,200 per capita in annual direct and indirect expenses (i.e., treatment-related and productivity costs; Ozminkowski et al., 2007). The burden of poor sleep increases with age (e.g., Gadie et al., 2017; Vitiello et al., 2004), for reasons that include increased sleep fragmentation (Ohayon et al., 2004) and psychiatric comorbidities (Brewster et al., 2018), and elevated insomnia rates have been observed among racial and ethnic minorities and women (e.g., Grandner, 2017).

Like insomnia, cigarette smoking is also highly prevalent and associated with substantial mortality and economic burdens, contributing to one-half million premature deaths and costing $289 billion each year (United States Department of Health and Human Services, 2014). Despite known health risks, including cardiovascular disease, chronic respiratory disease, and cancer (United States Department of Health and Human Services, 2014), approximately 14% of adults in the United States continue to smoke tobacco cigarettes (Clarke et al., 2018). There is also converging evidence that rates of insomnia are higher among smokers (e.g., 26%; Taylor et al., 2018) than in the general population (Chen et al., 2017; Ford et al., 2015; Lee et al., 2016; Purani et al., 2019; Taylor et al., 2018), and that individuals reporting insufficient sleep duration are more likely to be regular smokers (United States Centers for Disease Control and Prevention, 2016).

Several possible explanations for the observed high rates of insomnia among smokers have been put forth. First, nicotine is a stimulant that can drive changes in sleep architecture (e.g., longer sleep latency, shorter total sleep duration, and less slow wave sleep; Gibson et al., 2019; Zhang et al., 2006), and ultimately lead to the development of insomnia disorder among tobacco cigarette smokers (Fucito et al., 2014; Zhang et al., 2006). Second, approximately 19% of heavy smokers report waking during the night to smoke (Rieder et al., 2001), which may contribute to sleep disturbance (Peters et al., 2011). Third, insomnia may be a characteristic of nicotine withdrawal (e.g., Hughes, 2007). Abstinent smokers frequently report sleep disturbance in the context of withdrawal (e.g., Hughes, 2007; Peters et al., 2011). Even among non-abstaining smokers, brief abstinence that occurs during sleep each night has been shown to produce a drop in blood-nicotine levels that may trigger withdrawal symptoms and sleep fragmentation (Liao et al., 2019; Peters et al., 2011; Wetter & Young, 1994; Zhang et al., 2006), which is consistent with findings that smokers with higher levels of nicotine dependence exhibit greater sleep discontinuity (e.g., Cohen et al., 2020).

Identifying cognitive-affective mechanisms that underlie associations between smoking and insomnia is a critical next step to inform treatments for nicotine users with sleep disturbance. Accumulating evidence suggests that distress intolerance (i.e., the ability to withstand aversive emotional and/or physical states; Simons & Gaher, 2005) may contribute to both insomnia and smoking. For example, distress intolerance has been associated with poor sleep quality (Short et al., 2016) and greater number of years as a regular smoker (Leventhal & Zvolensky, 2015; Leyro et al., 2010), and is a reliable predictor of poorer smoking cessation outcomes among both community and treatment-seeking samples (e.g., Abrantes et al., 2008; Leventhal & Zvolensky, 2015; Steinberg et al., 2012; Veilleux, 2019).

Recent research has further underscored the utility of domain-specific subtypes of distress intolerance (McHugh & Otto, 2011; Sirota et al., 2013), such as intolerance for smoking abstinence (Sirota et al., 2010). Intolerance for smoking abstinence involves both intolerance for nicotine withdrawal symptoms and difficulty rationalizing the withdrawal experience (i.e., lack of cognitive withdrawal-coping tools; Rohsenow et al., 2015). Compared with general distress intolerance, intolerance for smoking abstinence is more strongly associated with detrimental smoking outcomes (higher smoking use rates, greater dependence, shorter past cessation attempts; Sirota et al., 2013; Sirota et al., 2010), and the two subscales have demonstrated unique associations with cessation outcomes (e.g., Lack of Cognitive Coping with fewer past serious quit attempts and Withdrawal Intolerance with prospective cessation failure; Rohsenow et al., 2015; Sirota et al., 2010). There is also reason to believe that intolerance for smoking abstinence may be associated with greater difficulty initiating and/or maintaining sleep among tobacco cigarette smokers. For example, given that smokers often report nighttime withdrawal symptoms (Liao et al., 2019; Peters et al., 2011; Wetter & Young, 1994), those who have greater difficulty tolerating the distress of withdrawal symptoms may also experience greater difficulty initiating/maintaining sleep.

The goal of this study was to conduct the first test of associations between intolerance for smoking abstinence and insomnia among a sample of daily cigarette smokers. We assessed two aspects of insomnia (impact and severity), given that impacts on functioning (e.g., daytime fatigue, distress, and lower work productivity) are primary reasons driving individuals with insomnia to seek help (Kyle et al., 2010; Morin et al., 2006). Specifically, we hypothesized that individuals who reported greater intolerance for nicotine withdrawal symptoms and lack of cognitive coping skills for withdrawal would be more likely to score above the threshold for clinically significant insomnia symptoms on a self-report measure (‘clinical insomnia’), and would also endorse greater insomnia symptom severity (i.e., difficulties initiating, maintaining, and transitioning out of sleep) and impact on functioning (i.e., the effects that insomnia symptoms have on distress and daytime functioning).

Method

Participants and Procedure

These are secondary analyses of baseline data collected during an experimental study of the effects of nicotine deprivation on experimental pain reactivity (Ditre et al., 2018). Participants were recruited from the community and were required to be between 18–65 years and currently smoking ≥ 15 cigarettes per day. Participants provided informed consent and smoking status was verified via exhaled carbon monoxide (CO ≥ 8ppm). Participants were excluded if they reported a current attempt to quit or reduce smoking, chronic pain, current use of prescription pain medications, or an inability to speak/read English. All activities were approved by the Institutional Review Board of Syracuse University (Protocol Ditre-12–228, ‘Smoking and Mood Study’).

Measures

Insomnia Severity Index.

The Insomnia Severity Index (ISI; Bastien et al., 2001) is a widely-used 7-item measure that evaluates severity of sleep disturbance. Items are rated on 5-point Likert-type scales ranging from 0 (‘not at all’) to 4 (‘extremely’), and higher total scores reflect more severe insomnia (range: 0 to 28). Thresholds for clinically significant insomnia have been established, with scores ≥ 15 indicating clinical insomnia (Lauriola et al., 2019; Morin et al., 2011). The ISI is also composed of two subscales (i.e., Impact and Severity). Three items measuring distress, interference with daily functioning, and noticeability of impairment are summed to yield an “impact on functioning” score (Impact subscale), and three items measuring severity of problems with sleep onset, sleep maintenance, and morning awakening are summed to yield a “symptom severity” score (Severity subscale; Bastien et al., 2001; Otte et al., 2019). Confirmatory factor analysis has supported a two-factor solution comprising severity and impact domains (Otte et al., 2019). Evaluation of insomnia impact and severity separately is further supported by their differential associations with clinical and demographic factors (e.g., impact but not severity demonstrating associations with depressive symptoms; Bazargan et al., 2019). The ISI has shown high internal consistency in both community and clinical samples (Cronbach’s α = .90, α = .91), and good convergent validity (large correlations with other self-report measures of sleep quality and depression, moderate correlations with fatigue; Morin et al., 2011). ISI scores (with a cut-off of 15) detected approximately 80% of cases with insomnia disorder in a psychiatric sample (Seow et al., 2018). In our sample, the ISI total, Severity, and Impact scales each demonstrated high internal consistency (Cronbach’s α = .92, α = .80, α = .90, respectively).

Intolerance for Smoking Abstinence Discomfort Questionnaire (IDQ-S).

The Intolerance for Smoking Abstinence Discomfort Questionnaire (IDQ-S; Sirota et al., 2010) is a 17-item measure designed to assess intolerance for symptoms of nicotine withdrawal. Items are rated on a Likert-type scale ranging from 1 (‘strongly disagree’) to 5 (‘strongly agree’). The Withdrawal Intolerance subscale score (range: 1 to 5, with 5 representing the highest intolerance for withdrawal symptoms) is the mean of 12 items that assess intolerance of cognitive, affective, and physiological symptoms associated with nicotine withdrawal (e.g., “I cannot stand how I feel when I need a cigarette”). The Lack of Cognitive Coping subscale score (range: 1 to 5, with 5 representing the least coping skills) is the mean of 5 items that assess an absence of cognitive tools used to tolerate withdrawal (e.g., “To get through a day without a cigarette, I think to myself, “no pain, no gain”; reverse coded). Both IDQ-S subscales have previously demonstrated high internal consistency, and have been associated with heavier cigarette use, greater dependence, and shorter length of longest past cessation attempt (Sirota et al., 2010). Moreover, the two subscales have demonstrated differential associations with cessation outcomes. For example, the Lack of Cognitive Coping subscale has been associated with fewer past serious quit attempts (Sirota et al., 2010), whereas the Withdrawal Intolerance subscale has prospectively predicted poorer cessation outcomes (Rohsenow et al., 2015). In the present sample, the IDQ-S Withdrawal Intolerance and IDQ-S - Lack of Cognitive Coping each demonstrated good to excellent internal consistency (Cronbach’s α = 0.93, α = .80, respectively).

Heaviness of Smoking Index.

Cigarette dependence was assessed using the Heaviness of Smoking Index (HSI; Heatherton et al., 1989), a widely used measure comprising two items: “How soon after you wake up do you smoke your first cigarette?” and “How many cigarettes per day do you smoke?”. Items are summed to create a total score (range: 0 to 6), with higher scores indicating greater cigarette dependence. HSI scores are significantly associated with biochemical indicators of cigarette use (e.g., exhaled carbon monoxide; Kozlowski et al., 1994), and predict success of cessation attempts (Breslau & Johnson, 2000; Burling & Burling, 2003; Etter et al., 1999). Due to previously documented associations with withdrawal intolerance (e.g., Sirota et al., 2010), HSI scores were included as an a priori covariate.

Distress Intolerance Index.

The Distress Intolerance Index (DII; McHugh & Otto, 2012) is a 10-item measure of self-reported distress intolerance, i.e., the perceived inability to withstand aversive emotional and/or physical states. Items are summed for a total score from 0 to 50, with higher scores indicating greater intolerance of distress. Greater DII scores have demonstrated positive associations with poorer performance in behavioral persistence tasks and higher drug craving among individuals with substance dependence in prior work (McHugh & Otto, 2011). Internal consistency in the present sample was good (Cronbach’s α = .82). Given that general distress intolerance has been positively associated with poorer sleep (Short et al., 2016) and with abstinence-specific withdrawal intolerance among tobacco smokers (e.g., Sirota et al., 2010), DII score was selected as an a priori covariate.

Sociodemographic Characteristics.

Participants self-reported gender, age, Hispanic ethnicity, race, marital status, educational attainment, and household income. Response options for gender consisted of “female” and “male.” Response options for race consisted of “American Indian/Alaska Native,” “Asian,” “Native Hawaiian or Other Pacific Islander,” “Black or African American,” and “White.” Given the small number of participants who identified as a race other than white or black, we grouped all non-white participants together to create a dichotomous white/non-white variable.

Data Analytic Strategy

All analyses were conducted using SPSS Statistics 26 (IBM Corp, 2019). First, data were evaluated for assumptions of linear and logistic regression. Distributions of ISI subscales, IDQ-S scores (total and subscales), and all continuous covariates were examined for normality, and both skewness and kurtosis values were acceptable (George & Mallery, 2011), and sample size exceeded that necessary to detect an effect (~105) that may be characterized as small in magnitude (Cohen, 1992). Second, two separate hierarchical logistic regression models were conducted to test associations between the two IDQ-S subscale scores (Withdrawal Intolerance and Lack of Cognitive Coping) and likelihood of scoring above the cut-off for clinical insomnia (ISI total score ≥ 15). Third, two separate hierarchical linear regression models were conducted to test associations between the two IDQ-S subscale scores and (a) ISI Severity and (b) ISI Impact subscale scores. Age, gender, race, nicotine dependence, and general distress intolerance were included as covariates in all models, given previously observed relations with insomnia symptoms (e.g., Gadie et al., 2017; Grandner, 2017; Ohayon, 2002; Peters et al., 2011; Short et al., 2016). In each model, predictors were entered in the following order: Step 1 (age, race, gender, HSI scores, and DII scores); Step 2 (IDQ-S subscale score). Each model was also run in the reverse direction. Specifically, hierarchical linear regression models were conducted to test associations of (a) ISI Severity scores, (b) ISI Impact scores, and (c) clinically significant insomnia, with scores on the two IDQ-S subscales.

Results

Participant Characteristics

Participants included 224 current daily tobacco smokers (42.9% female; M_age = 41.5, SD = 12.3; 58.5% white), who smoked an average of 21 cigarettes per day (M = 21.3, SD = 11.1) and had been regular smokers for over 24 years (M = 24.4, SD = 12.2). The mean HSI score was 3.8 (SD = 1.3), corresponding to a high level of dependence (Schnoll et al., 2013). Over half of all participants (58.0%) reported 12 years of education or less. Over one-third (36.7%) of participants scored above the cut-off for clinical insomnia (ISI score ≥ 15). Additional sample characteristics are displayed in Table 1.

Table 1.

Participant Characteristics

	n (%)
Gender
Female	96 (42.9)
Ethnicity
Hispanic/Latino	9 (4.0)
Race
White	131 (58.5)
Black or African American	86 (38.4)
American Indian/Alaska Native	7 (3.1)
Marital Status
Single	135 (60.3)
Married	35 (15.6)
Divorced/Separated/Widowed	54 (24.1)
Education
0–11 Years	50 (22.3)
12 Years	80 (35.7)
12–15 Years	80 (35.7)
≥16 Years	14 (6.3)
Household Income
<$10,000	87 (38.8)
$10,000–29,999	79 (35.3)
$30,000–49,999	27 (12.1)
≥$50,000	31 (13.8)
Anxiety Symptoms a
None to mild	136 (60.7)
Moderate to severe	87 (38.8)
Hazardous or Harmful Drinking	70 (31.3)
Past-6-Month Cannabis Use	97 (43.3)
	M (SD)	Range
Age	41.5 (12.3)	18 – 65
Cigarettes Per Day	21.3 (11.1)	3 – 60
Years of Regular/Daily Smoking	24.4 (12.2)	3 – 55
HSI Cigarette Dependence b	3.8 (1.3)	0 – 6
ISI b total score	11.5 (7.6)	0 – 28
ISI Severity score	4.6 (3.4)	0 – 12
ISI Impact score	4.5 (3.4)	0 – 12

N = 224.

^aGeneral Anxiety Disorders – 7

^bHeaviness of Smoking Index

^cInsomnia Severity Index

Intolerance for Smoking Abstinence and Clinical Insomnia

IDQ-S – Withdrawal Intolerance scores were positively associated with likelihood of scoring above the ISI cut-off for clinical insomnia (adjusted odds ratio [AOR] = 1.95, 95% confidence interval [CI]: 1.29 – 2.92, P = .001; Table 2). Specifically, for every one-point increase in IDQ-S – Withdrawal Intolerance score, participants were nearly twice as likely to score above the cut-off for clinical insomnia. IDQ-S – Lack of Cognitive Coping scores were not associated with the likelihood of scoring above the ISI cut-off for clinical insomnia (AOR = .95, 95% CI: .66 – 1.38, P = .790; Table 2). The Nagelkerke pseudo-R² estimate for the combined model was .137 (Nagelkerke, 1991). In the reverse models, greater likelihood of endorsing clinically significant insomnia symptoms was associated with greater IDQ-S – Withdrawal Intolerance (β = .220, P = .001, ΔR² = .046) but not Lack of Cognitive Coping (β = −.058, P = .403, ΔR² = .003; Appendix Table A).

Table 2.

Logistic Regression: Likelihood of Clinical Insomnia as a Function of Intolerance for Smoking Abstinence

Variable	B	SE	AOR	95% CI	p
Gender	.422	.305	1.526	.839 – 2.775	.166
Age	−.012	.012	.988	.965 – 1.012	.328
Race	−.031	.304	.969	.534 – 1.760	.918
HSIa	.149	.125	1.161	.909 – 1.482	.233
DIIb	.005	.019	1.005	.967 – 1.043	.816
IDQ-Sc Lack – of Cognitive Coping	−.051	.190	.951	.655 – 1.379	.790
IDQ-Sc – Withdrawal Intolerance	.665	.208	1.945	1.294 – 2.923	.001

Note: N = 224. Results shown are from the second step of the logistic regression model; AOR = adjusted odds ratio; Race: Reference group = white; Gender: Reference group = female

^aHeaviness of Smoking Index

^bDistress Intolerance Index

^cIntolerance for Smoking Abstinence Discomfort Questionnaire

Intolerance for Smoking Abstinence and Insomnia Severity/Impact

Overall models predicting ISI Severity and ISI Impact were each significant (F[1,216] = 16.413, p < .001, R² = .129; F[1,216] = 34.561, p < .001, R² = .220). IDQ-S – Withdrawal Intolerance scores were positively associated with scores on both ISI Severity (β = .283, P < .001, ΔR² = .066) and Impact subscales (β = .389, P < .001, ΔR² = .125; Table 3). IDQ-S – Withdrawal Intolerance scores accounted for nearly 7% of the unique variance in symptom severity and 13% of the unique variance in impact on functioning, after accounting for all other variables in the models. No associations were observed between the IDQ-S – Lack of Cognitive Coping subscale and scores on either ISI subscale (ps > .43; Table 3). In the reverse models, ISI Severity and Impact were positively associated with IDQ-S – Withdrawal Intolerance (β = .269, P < .001, ΔR² = .068; β = .362, P < .001, ΔR² = .119), but not Lack of Cognitive Coping (β = − .102, P = .138, ΔR² = .010; β = −.036, P = .609, ΔR² = .001; Appendix Table B).

Table 3.

IDQ-S Subscales on Insomnia Severity Index Subscale Scores

	ISI Severity subscale			ISI Impact subscale
Variable	β	t	p	β	t	p
Gender	−.006	−1.018	.310	.037	.603	.547
Age	.024	.379	.705	−.043	−.704	.482
Race	−.005	−.074	.941	.062	1.012	.313
HSIa	.114	1.663	.098	.117	1.810	.072
DIIb	.047	.713	.477	.095	1.511	.132
IDQ-Sc –Lack of Cognitive Coping	−.038	−.580	.563	.050	.798	.426
IDQ-Sc –Withdrawal Intolerance	.283	4.051	<.001	.389	5.879	<.001
R 2		.129			.220
ΔR2		.066			.125
F for ΔR2		16.413**			34.561**

Note: N = 224. Results shown are from the second step of each linear regression model; β = standardized beta weights; Race: Reference group = white; Gender: Reference group = female

^aHeaviness of Smoking Index

^bDistress Intolerance Index

^cIntolerance for Smoking Abstinence Discomfort Questionnaire

^*p < .05

^**p < .01

Discussion

This is the first study to test associations between intolerance for smoking abstinence and insomnia severity. Results indicated that withdrawal intolerance was positively associated with insomnia symptoms, including both symptom severity and impact on functioning, among a sample of daily tobacco cigarette smokers, after controlling for relevant sociodemographic and cognitive-affective variables (age, gender, race, nicotine dependence, and general distress intolerance). The Withdrawal Intolerance component of the IDQ-S reflects inability to withstand the cognitive, affective, and physiological symptoms associated with nicotine withdrawal. Moreover, each one-point increase in IDQ-S – Withdrawal Intolerance subscale score was associated with a nearly two-fold greater likelihood of scoring above the cut-off for clinical insomnia, even after accounting for general distress intolerance. Scores on the IDQ-S – Lack of Cognitive Coping subscale were not associated with insomnia symptom severity or impact on functioning. Taken together, these findings suggest that smokers who are less confident in their ability to withstand smoking abstinence/withdrawal may also report greater sleep disturbance and related impairment (e.g., interference with daytime functioning).

Intolerance for withdrawal symptoms may be one important factor that contributes to increased rates of insomnia observed among cigarette smokers. Previous work indicates high rates of comorbidity between insomnia and tobacco dependence (Chen et al., 2017; Lee et al., 2016; Purani et al., 2019; Taylor et al., 2018), and researchers have suggested that nicotine withdrawal may contribute to insomnia among smokers (Liao et al., 2019; Peters et al., 2011; Wetter & Young, 1994). Although intolerance for smoking abstinence has been investigated as a predictor of cessation outcomes (e.g., Germeroth et al., 2018; Kahler et al., 2013), this is the first analysis to examine its relation to insomnia. These findings also build upon previous evidence that general distress intolerance is associated with poor sleep (e.g., Short et al., 2016) by investigating a domain-specific distress intolerance measure that may have particular relevance for treatment of smokers with comorbid insomnia. Other domain-specific measures of distress intolerance (e.g., Frustration Discomfort Scale; Harrington, 2005; Discomfort Intolerance Scale; Schmidt et al., 2006) have demonstrated unique associations with criterion variables (McHugh et al., 2011), and intolerance for withdrawal symptoms has shown greater utility in predicting smoking cessation outcomes than more general measures of distress intolerance (McHugh & Otto, 2011; Sirota et al., 2013). In the present sample, intolerance for withdrawal symptoms accounted for significant additional variance in insomnia severity above and beyond that accounted for by general distress intolerance.

These results may further be understood within a broader empirical literature suggesting that cognitive-affective responses play critical roles in sleep disturbance (Espie, 2002; Tang, 2009; Tang et al., 2012). For example, in accordance with Espie’s model of insomnia (Espie, 2002), pre-sleep cognitive processes (especially affect-laden thoughts) have been suggested to interrupt the automatic initiation of sleep and to produce sleep disturbance. Among smokers, pre-sleep cognitive arousal related to the experience of early withdrawal (e.g., via intolerance for withdrawal symptoms) may contribute to difficulty initiating sleep. Future research could examine this by evaluating associations between measures of pre-sleep cognitive arousal (e.g., the Pre-Sleep Arousal Scale; Nicassio et al., 1985) and intolerance for smoking abstinence, and by assessing cognitive processes among smokers during the period of nightly sleep initiation (perhaps using self-report indices of nighttime thought content or audio-recordings of spontaneous thoughts; Fichten et al., 1998; Wicklow & Espie, 2000).

Our hypothesis that IDQ-S – Lack of Cognitive Coping subscale scores would be positively associated with insomnia severity/impact was not supported. Previous work has demonstrated that scores on this subscale are negatively associated with number of past quit attempts (i.e., greater coping skills are associated with more attempts; Sirota et al., 2013), but not with cessation failure among treatment-seeking smokers (Rohsenow et al., 2015). Indeed, researchers have suggested that cognitive skills for coping with nicotine withdrawal are more relevant in some withdrawal situations (e.g., unassisted quit attempts) than others (Rohsenow et al., 2015), and one possibility is that smokers are less likely to engage cognitive coping strategies when tolerating incidental withdrawal symptoms (i.e., during nightly sleep) than when actively attempting to quit and/or during periods of extended abstinence. Future work should examine associations between cognitive coping strategies for withdrawal and insomnia symptoms among smokers who are engaging in a quit attempt or have been deprived from cigarettes for longer periods (e.g., > 12 hours). In addition, the IDQ-S assesses only cognitive coping strategies, and future work could examine the role of other coping strategies (e.g., behavioral strategies) for tolerating withdrawal. Future research is needed to replicate these findings and clarify the relative importance of intolerance for nicotine withdrawal and availability of cognitive withdrawal-coping tools in relation to insomnia among individuals with comorbid nicotine dependence.

Given prior evidence that smoking and insomnia may exacerbate each other in a bidirectional manner (e.g., Peters et al., 2011; Purani et al., 2019; Short et al., 2016), one potential clinical implication of this and future research is the possibility that smokers with insomnia may benefit from interventions that specifically address intolerance for smoking withdrawal. Emerging evidence suggests that cognitive behavioral therapy (CBT) can reduce insomnia symptoms in treatment-seeking smokers (Fucito et al., 2014; Short et al., 2017), and such interventions could be adapted to also address intolerance for smoking abstinence. It is also possible that regular smokers may benefit from assessment of intolerance for smoking abstinence and targeted treatments to improve sleep outcomes. For example, several cognitive-behavioral and third-wave insomnia therapies focus on building skills to tolerate distress (Taylor et al., 2015), and integrated approaches may benefit from including components that challenge dysfunctional thoughts about the consequences of smoking abstinence.

Several limitations of this study should be noted. First, these cross-sectional analyses preclude causal interpretation and inferences regarding directionality or temporal precedence. For example, although intolerance for withdrawal symptoms could contribute to difficulty initiating sleep among smokers, it is also possible that insomnia symptoms reduce ability to tolerate withdrawal, and that use of nicotine to regulate arousal following episodes of poor sleep could exacerbate nicotine dependence and intolerance for withdrawal (Gibson et al., 2019; Taylor et al., 2018). Future research should examine covariation between trajectories of insomnia and intolerance for smoking abstinence, both over time and in the context of a quit attempt, perhaps using ecological momentary assessment (e.g., Cox et al., 2018; Langdon et al., 2016; Vinci et al., 2018). Second, the assessment of intolerance for smoking abstinence was not specific to nighttime or pre-sleep withdrawal experience, and future research could evaluate whether pre-sleep thought content is related to early smoking abstinence withdrawal symptoms (e.g., Hendricks et al., 2006) among smokers with insomnia. Third, participants in the current study were heavy smokers with high levels of nicotine dependence, and relatively low educational attainment and household income. Thus, the generalizability of these findings to light or intermittent smokers and groups with different socioeconomic status remains unclear. This work should also be extended to smokers who co-use other nicotine products (e.g., e-cigarettes). Future work should include comprehensive assessment of the role of other substances (e.g., alcohol, cannabis, and sleeping medications/sedatives) in withdrawal intolerance-insomnia relations. Fourth, we did not assess whether withdrawal intolerance might moderate relationships of insomnia with nicotine dependence, and note this as a direction for future work. Finally, although an ISI cut-off of 15 detects approximately 80% of cases with insomnia disorder (Seow et al., 2018), future research should incorporate more comprehensive assessment of diagnostic criteria (e.g., Structured Clinical Interview for DSM-5; First et al., 2015).

In summary, these data provide initial evidence of positive associations between intolerance for nicotine withdrawal and insomnia symptoms/severity. Although intolerance for withdrawal may contribute to the observed higher rates of sleep disturbance among tobacco cigarette smokers (relative to non-smokers), prospective research is needed to examine covariation and temporal effects between intolerance for withdrawal and insomnia.

Public health significance:

Tobacco cigarette smokers who report greater difficulty tolerating nicotine withdrawal symptoms may also report more severe insomnia. These findings suggest that intolerance for withdrawal may contribute to high rates of insomnia among cigarette smokers, and future work should investigate the clinical utility of teaching smokers with insomnia to better cope with withdrawal symptoms.

Appendix Table A.

Clinically Significant Insomnia Scores (ISI > 14) on IDQ-S Subscales

	IDQ-Sc –Lack of Cognitive Coping			IDQ-Sc –Withdrawal Intolerance
Variable	β	t	p	β	t	p
Gender	−.009	−.127	.899	−.155	−2.464	.015
Age	.002	.030	.976	−.110	−1.773	.078
Race	−.031	−.454	.650	−.071	−1.139	.256
HSIa	.136	1.955	.052	.202	3.158	.002
DIIb	−.065	−.929	.354	.116	1.810	.072
Clinically significant ISI	−.058	−.837	.403	.220	3.484	.001
R 2		.022			.176
ΔR2		.003			.046
F for ΔR2		.701			12.137**

Note: N = 224. Results shown are from the second step of each linear regression model; p = standardized beta weights; Race: Reference group = white; Gender: Reference group = female; Clinically significant ISI: Reference group = no clinical insomnia

^aHeaviness of Smoking Index

^bDistress Intolerance Index

^cIntolerance for Smoking Abstinence Discomfort Questionnaire

^*p < .05

^**p < .01

Appendix Table B.

Insomnia Severity Index Subscale Scores on IDQ-S Subscales

	IDQ-Sc –Lack of Cognitive Coping			IDQ-Sc –Withdrawal Intolerance
Variable	β	t	p	β	t	p
Gender	−.013	−.191	.849	−.153	−2.473	.014
Age	.006	.093	.926	−.128	−2.095	.037
Race	−.033	−.486	.628	−.066	−1.074	.284
HSIa	.146	2.091	.038	.186	2.936	.004
DIIb	−.058	−.835	.404	.101	1.597	.112
ISI Severity subscale	−.102	−1.487	.138	.269	4.301	< .001
R2		.029			.198
ΔR2		.010			.068
F for ΔR2		2.211			18.501**
Gender	−.002	−.031	.975	−.172	−2.898	.004
Age	.004	.062	.951	−.098	−1.642	.102
Race	−.029	−.430	.668	−.084	−1.413	.159
HSIa	.136	1.921	.056	.156	2.514	.013
DIIb	−.062	−.878	.381	.073	1.188	.236
ISI Impact subscale	−.036	−.512	.609	.362	5.855	< .001
R 2		.020			.249
ΔR2		.001			.119
F for ΔR2		.262			34.284**

Note: N = 224. Results shown are from the second step of each linear regression model; β = standardized beta weights; Race: Reference group = white; Gender: Reference group = female

^aHeaviness of Smoking Index

^bDistress Intolerance Index

^cIntolerance for Smoking Abstinence Discomfort Questionnaire

^*p < .05

^**p < .01