This cross-sectional study used data from the National Health and Nutrition Examination Survey (NHANES), a nationally representative survey conducted by the Centers for Disease Control and Prevention (CDC) in the United States. NHANES employs a complex, multistage probability sampling design to assess the health and nutritional status of the civilian, non-institutionalized U.S. population (15).

Data from six consecutive NHANES cycles (2007–2008, 2009–2010, 2011–2012, 2013–2014, 2015–2016, and 2017- 2020 March) were combined for this study, yielding a total sample size of 66,148 participants. Among these, 30,269 subjects had complete data on smoking status (including non-smokers) and were included in the analysis. Subjects who did not provide information on their smoking status (n = 35,879) were excluded.

The study population varied depending on the sleep outcome of interest. For the analysis of sleep duration, 22,745 subjects with valid sleep duration data were included. The analysis of sleep problems included 22,796 subjects with available data. For the analyses of snoring, snorting or stop breathing during sleep, and overly sleepy, 10,905, 10,487, and 10,099 subjects were included, respectively, based on the availability of relevant data ( Figure 1 ). Basic characteristics of included and excluded participants, including age, BMI, gender and race distribution were compared ( Supplementary Table S1 ).

All participants provided written informed consent, and the NHANES study protocol was approved by the National Center for Health Statistics Research Ethics Review Board (15).

Smoking status was determined using the NHANES questionnaire item “SMQ020 - Smoked at least 100 cigarettes in life”. Participants who reported having smoked less than 100 cigarettes in their lifetime were classified as “non-smokers”. Among those who had smoked at least 100 cigarettes, we define the previous smoker by determining subjects whose answer for item “SMQ040 - Do you now smoke cigarettes?” is “3, Not at all”. For participants currently smoking, the intensity of smoking was assessed using the questionnaire item “SMD650 - Avg # cigarettes/day during past 30 days”.

Based on the responses to SMD650, smokers were further categorized into three groups: light smokers, who reported smoking more than 0 but fewer than 5 cigarettes per day, moderate smokers, who reported smoking more than 5 but fewer than 10 cigarettes per day and heavy smokers, who reported smoking 10 or more cigarettes per day.

For the dose-response analysis, the continuous variable derived from “SMD650 - Avg # cigarettes/day during past 30 days” was used to represent the average number of cigarettes smoked per day. This approach allowed for a more granular examination of the relationship between smoking quantity and sleep outcomes.

This study examined several sleep-related outcomes, including sleep duration sufficiency, diagnosed sleep problems, the frequency of snoring or snorting and stop breathing during sleep, and feeling overly sleepy during the day.

Sleep duration was assessed using two different questionnaire items across the NHANES cycles. In the 2007–2014 cycles, sleep duration was measured using the item “SLD010H - How much sleep do you get (hours)?”. For the 2015–2020 cycles, sleep duration was assessed using the item “SLD012 - Sleep hours”. Based on previous research, sleep duration was dichotomized into sufficient (>7 hours per night) and insufficient (≤7 hours per night) sleep (6, 14). Diagnosed sleep problems were determined using the questionnaire item “SLQ050 - Ever told doctor had trouble sleeping?”. Participants who reported being told by a doctor that they had trouble sleeping were considered to have a diagnosed sleep problem.

Snoring frequency was assessed using the questionnaire item “SLQ030 - How often do you snore?”. Responses were categorized as ‘never’, ‘rarely’ (1–2 nights per week), ‘occasionally’ (3–4 nights per week), and ‘frequently’ (5 or more nights per week).

The frequency of snorting or stop breathing during sleep was measured using the questionnaire item “SLQ040 - How often do you snort or stop breathing?”. Responses were categorized as ‘never’, ‘rarely’ (1–2 nights per week), ‘occasionally’ (3–4 nights per week), and ‘frequently’ (5 or more nights per week).

The frequency of feeling overly sleepy during the day was defined by “SLQ120 - How often feel overly sleepy during day?”, and Responses were categorized as ‘never’, ‘rarely’ (1 time a month), ‘sometimes’ (2–4 times a month), ‘frequently’ (5–15 times a month), and ‘almost always’ (more than 15 times a month).

In the dose-response analysis, we treated the ordinal categorical variables (snoring, snorting or stop breathing, and feeling overly sleepy) as binary variables. We defined ‘never’ and ‘rarely’ as 0, and ‘occasionally’ or more frequent categories as 1. The purpose of this approach was to investigate the relationship between smoking and the presence of occasional or more severe sleep problems, which we believe has more practical significance.

Covariables were chosen based on previous research (13, 14). Demographic characteristics were extracted from the demographic questionnaire, including age, gender, race/ethnicity, marital status, family poverty income ratio, and education level. These variables were categorized as follows: Race/ethnicity: Non-Hispanic White, Non-Hispanic Black, Mexican American, Other Hispanic, and Other Race - Including Multi-Racial. Marital status: Married/living with partner, Widowed/divorced/separated, and Never married. Family poverty income ratio: Low income (<1), Middle income (1–3), and High income (≥3). Education level: Below high school, High school/GED, and College or above. Body mass index (BMI) was extracted from “BMXBMI - Body Mass Index (kg/m**2)”. BMI was treated as a continuous variable in the analysis. The exercise load is determined based on whether people have done 75–150 minutes of high-intensity exercise or 150–300 minutes of moderate-intensity exercise per week based on the physical activity guidelines for American (16). Subjects were categorized into lack of exercise (below the standard), adequate exercise (meeting the standard), or highly active exercise (exceeding the standard). Dietary energy intake was also considered. The recommended energy intake for males is 2000–3000 kcal per day, while for females, it is 1600–2400 kcal per day (17). Based on these recommendations, subjects were divided into three categories: low energy intake, adequate energy intake, or excessive energy intake. Medication status, including the presence of asthma, diabetes, heart failure, emphysema, and chronic bronchitis, was also considered as a covariate.

According to the NHANES protocol, all the data were integrated into a single dataset, and data analysis took into account the masked variance and applied the suggested weighting methodology. Sample weights from the Mobile Examination Center (MEC) interviews were used to address non-response, non-coverage, and unequal probabilities of selection (18).

Participants were divided into two groups based on sleep sufficiency or reported sleep problems, or divided into groups based on the frequency of snoring, snorting, or stopping breathing during sleep, or being overly sleepy during the day. To explore the differences between these groups, weighted descriptive analyses were performed. For continuous variables such as age and BMI, weighted percentages were computed. For categorical variables such as gender, race, education level, and smoking status, weighted percentages were computed, and the weighted chi-square test was used to assess group differences. Data are expressed as mean ± SE. Odds ratios (ORs) and 95% confidence intervals (CIs) for the association between sleep disorders and smoking status were calculated using weighted logistic regression models. Both unadjusted and multivariate adjusted models were used in this research: the crude model was adjusted for no covariates; Model 1 was adjusted for age, sex, and race/ethnicity; Model 2 was further adjusted for body mass index, education level, marital status, poverty level, exercise level, dietary energy intake level, and medication status.

To determine the association of smoking with sleep disorders, we performed the Cochran-Armitage trend test and used logistic regression models to test whether sleep disorder outcomes exhibited a positive trend across smoking levels, while adjusting for potential confounders.

The non-linear relationship between the amount of smoking and binary sleep outcomes was evaluated using restricted cubic splines. A dose-response curve was plotted to visualize the relationship, with the odds ratios and 95% CIs predicted for different smoking amounts using the Predict function.

All statistical analyses were performed using R software version 4.3.2 (http://www.R-project.org, The R Foundation). A p-value of < 0.05 was statistically significant.

Table 6 presents the weighted associations between smoking status and various sleep disorders outcomes, including insufficient sleep duration, reported sleep problems, snoring, snorting or stopping breathing during sleep, and feeling overly sleepy, after adjusting for confounders. Crude models and models adjusted for age, gender, and race are provided in the Supplementary Data ( Supplementary Tables S2–6 ).

Compared to non-smokers, former smokers had significantly higher odds of reporting sleep problems (OR: 1.435, 95% CI: 1.310–1.573, P<0.001). Light smokers exhibited significantly higher odds of reporting sleep problems (OR: 1.364, 95% CI: 1.113–1.672, P=0.004) compared to non-smokers. Moderate smokers had significantly higher odds of insufficient sleep duration (OR: 1.272, 95% CI: 1.047–1.546, P=0.017) and reported sleep problems (OR: 1.749, 95% CI: 1.398–2.186, P<0.001) compared to non-smokers. Heavy smokers consistently demonstrated significant associations across multiple sleep outcomes. They had higher odds of insufficient sleep duration (OR: 1.732, 95% CI: 1.528–1.963, P<0.001), reported sleep problems (OR: 1.990, 95% CI: 1.766–2.243, P<0.001), snoring (OR: 1.908, 95% CI: 1.164–3.128, P=0.03), and snorting or stopping breathing during sleep (OR: 1.863, 95% CI: 1.183–2.936, P=0.022) compared to non-smokers. However, no significant association was found for feeling overly sleepy.

Table 7 presents the associations between smoking trend and various sleep-related outcomes. Compared to non-smokers, smokers had a higher odds of reporting insufficient sleep duration (OR: 1.112, 95% CI: 1.088–1.136, P<0.001). Similarly, smokers were more likely to report sleep problems (OR: 1.187, 95% CI: 1.159–1.215, P<0.001) and snoring (OR: 1.154, 95% CI: 1.114–1.195, P<0.001) than non-smokers. The odds of experiencing snorting or stopping breathing during sleep were also higher among smokers (OR: 1.172, 95% CI: 1.116–1.231, P<0.001). Additionally, smokers had a higher likelihood of feeling overly sleepy during the day compared to non-smokers (OR: 1.079, 95% CI: 1.043–1.116, P<0.001).

Figure 2 presents dose-response curves showing an increase in the adjusted odds ratios for various sleep outcomes, including insufficient sleep, reported sleep problems, snoring, snorting or stopping breathing, and feeling overly sleepy during the day, as the amount of smoking rises. The dose-response curves reveal that when the daily smoking amount is less than approximate 30 cigarettes, the risk of these sleep problems increases with the rise in smoking quantity, suggesting a dose-dependent adverse effect of smoking on sleep. However, when the daily smoking amount exceeds approximate 30 cigarettes, the curves exhibit a flattening or declining trend, accompanied by widening confidence intervals. This might be attributed to the relatively smaller sample size within the high smoking quantity range, leading to increased uncertainty in the estimates. The expanding confidence intervals suggest greater variability in the odds ratio estimates at higher smoking levels, but the overall patterns remain consistent.