The NHANES data from 2005 to 2010 (13) were used for this cross-sectional study conducted by the Centers for Disease Control and Prevention. Using a survey conducted in stages with varying levels of sampling probabilities, the NHANES project evaluated the health and nutritional situation of non-institutionalized Americans (14). Using a mobile examination center (MEC), an integrated system of home visits, laboratory tests, and screenings is applied to gather demographic and health information for the NHANES. Participants in the NHANES provided written informed consent before participating in the study, which was approved by the Ethics Review Committee of the National Center for Health Statistics (NCHS). It was unnecessary to obtain additional approval from the Institutional Review Board for the secondary analysis (15). Data from the NHANES can be accessed by visiting the NHANES website.¹ The research involved individuals aged 20 years and more who had completed the questionnaire. Individuals with missing serum cotinine levels, smoking status, or constipation were excluded.

The term “never smoker” is applied to smokers aged ≥20 who claim to have never smoked more than 100 cigarettes throughout their existence during the household interview. Individuals who were former smokers had smoked over 100 cigarettes before quitting, while those who are currently smoking smoke regularly. A measure of secondhand smoke exposure and active smoking is cotinine, nicotine’s primary metabolite (16). Isotope dilution-high-performance liquid chromatography/atmospheric pressure chemical ionization tandem mass spectrometry was used to measure serum cotinine levels (17). Previous investigations (18) considered those below the lower detection threshold unexposed. To represent smoking exposure, we created cotinine categories using the newly recommended cutoff point of 3 ng/mL, as suggested by a previous study (19), to differentiate smokers from non-smokers. Cotinine levels were categorized into three groups: cotinine <0.05 ng/mL, cotinine 0.05–2.99 ng/mL, and cotinine ≥3 ng/mL.

To diagnose constipation, stool frequency and consistency were considered important factors based on the Rome IV criteria (20) and prior NHANES studies (21–26). Our study defined chronic constipation primarily based on stool frequency, as assessed in the NHANES Intestinal Health Questionnaire. By asking about the frequency of stool habits, the frequency of bowel movements was determined: “How many times per week do you usually have a bowel movement?” Constipation (<3 stools/week), diarrhea (≥21 stools/week), and normal were classified as the results. Additionally, a card with colored photos and clarifications of the seven Bristol Stool Form Scale (BSFS) types (Type 1 to Type 7) was shown to the participants, and they were asked to indicate which type of stool corresponded to their most common stool type. Constipation is characterized by hard lumpy or lumpy stools of BSFS Type 1 (like nuts) or Type 2 (like sausages), and diarrhea is characterized by soft fragments with uneven borders and mushy stool types 6 or 7 (watery with no solid chunks). The remaining subjects were considered to have normal stool habits.

In accordance with the literature, a variety of potential covariates were assessed (3, 20, 21, 23, 26), including age, sex, ethnicity, education level, marital status, physical activity, body mass index (BMI), dietary factors [calorie consumption, protein consumption, carbohydrate consumption, sugar consumption, fiber consumption, fat consumption, caffeine consumption, alcohol consumption, water (moisture) consumption, and dietary supplements], and previous diseases (diabetes, stroke, and coronary heart disease). Ethnicity was categorized as non-Hispanic white, non-Hispanic black, Mexican American, or other ethnicities. Educational attainment was ranked as less than 9 years, 9–12 years, and more than 12 years. Marital status was classified as married, living with a partner, or living alone. Physical activity was defined as participating in high-intensity sports, exercise, or leisure activities, such as running or basketball, that cause a significant increase in breathing or heart rate for a continuous period of at least 10 min. According to a standardized technique, BMI was calculated using height and weight. Participants’ 24-h nutritional information was obtained via a dietary recall interview conducted before an interview at an MEC. This included data on total dietary calories, protein, carbohydrates, sugar, fiber, fat, caffeine, alcohol, and water (moisture). Supplement usage over the past month was determined through a question about nutritional supplements and medications. According to the questionnaire, previous diseases such as diabetes, stroke, and coronary heart disease, are based on whether a doctor had previously diagnosed the condition.

The means (standard deviations) and medians (interquartile ranges) were calculated and compared for continuous variables using either the Student’s t-test for normal distribution or the Mann–Whitney U-test for non-normal distribution. Categorical variables were displayed as absolute values in percentages and analyzed using the chi-square test. Cotinine levels were log2-transformed into a normal distribution as continuous variables.

We undertook several diagnostic tests to examine basic logistic regression assumptions (28), and then three consecutive multivariate logistic regression models were used to examine the relationship between tobacco exposure and the occurrence of constipation. Model 1 was adjusted for age, sex, ethnicity (non-Hispanic white, non-Hispanic black, Mexican American, or other ethnicities), educational attainment (less than 9 years, 9–12 years, and more than 12 years), physical activity, marital status (married, living with a partner, or living alone), and BMI (continuous). Model 2 was based on model 1 and added previous diseases (diabetes, stroke, and coronary heart disease). Model 3 was adjusted for model 2 and included additional adjustments for dietary factors [calorie consumption, protein consumption, carbohydrate consumption, sugar consumption, fiber consumption, fat consumption, caffeine consumption, alcohol consumption, and water (moisture) consumption] and dietary supplements.

In addition, restricted cubic spline (RCS) regression was used to examine linearity and assess the dose–response curve between log2-transformed serum cotinine levels and constipation risk after adjusting variables in model 3. Subgroup analyses were stratified by relevant confounders based on their association with the outcomes of interest. The analysis was conducted using R statistical software (Version 4.2.2, The R Foundation)² and Free Statistics analysis platform (Version 1.8, Beijing, China) (27, 29). All participants were involved in a descriptive study. Using two-tailed testing, we determined that a p-value of 0.05 was significant.

This study selected 31,034 potential participants from the NHANES (2005–2010). Excluded from the study were individuals under the age of 20 (n = 13,902) and those who lacked information on serum cotinine levels, smoking status, and constipation (n = 5,008). We included 12,124 participants after excluding those with missing covariate data (n = 473) in our analysis. The flowchart in Figure 1 illustrates the exclusion criteria. The baseline characteristics of the research individuals are shown in Table 1 according to the NHANES 2005–2010. In this sample, a total of 3,173 (27.2%) participants had cotinine levels ≥3 ng/mL, and 2,550 (21.9%) individuals in the reference group were current smokers. Compared to women, men have significantly higher cotinine levels.

The average age of the study population was 48.8 ± 17.9 years. In this study, 3,173 (27.2%) participants had cotinine levels above 3 ng/mL, 2,871 (24.6%) participants had cotinine levels between 0.05 and 2.99 ng/mL, and 5,607 (48.1%) participants had cotinine levels below 0.05 ng/mL. There were 2,550 (21.9%) current smokers, 2,914 (25.0%) former smokers, and 6,187 (53.1%) non-smokers among the participants who self-reported their smoking status. Self-reported current smokers and individuals with cotinine levels >3 ng/mL were more likely to be younger, male, and non-Hispanic white, have 9–12 years educational attainment, married or living with a partner, have less physical activity and BMI, have higher daily dietary intakes of energy, protein, carbohydrate, sugar, fat, caffeine, alcohol, moisture, lower fiber, and dietary supplements, and have a lower incidence of diabetes, coronary heart disease, and stroke than those with a serum cotinine level < 0.05 ng/ mL and those who were self-proclaimed non-smokers. We found that 3.8% of American adults experienced constipation based on bowel frequency.

Based on the univariate analysis, gender, age, ethnicity, marital status, physical activity, BMI, stroke, and dietary intakes, such as total energy, protein, carbohydrate, fiber, fat, alcohol, moisture, and dietary supplements were associated with constipation (Table 2). Using both continuous and categorical perspectives, the multivariate logistic regression analysis illustrates the relationship between serum cotinine levels, smoking status, and the risk of constipation (Table 3 and Supplementary Table 2). In contrast to our previous assumptions, after controlling for covariates, there was no significant relationship between the cotinine levels transformed by log2 and constipation risk. Interestingly, according to the categorical analysis, a significant effect of serum cotinine on stool frequency-related constipation is observed in a population with cotinine levels between 0.05 and 2.99 ng/mL. The incidence of constipation was 36% higher in participants with serum cotinine levels between 0.05 and 2.99 ng/mL (model 1: OR = 1.45 [1.13–1.85]; model 2: OR = 1.44 [1.12–1.83]; model 3: OR = 1.36 [1.06–1.74]; p < 0.05) after adjusting for covariates using the lowest levels of serum cotinine (<0.05 ng/mL) as a reference.

To determine the relationship between log2-transformed serum cotinine levels and constipation, restricted cubic spine regression with multivariable adjustment was used. There was a linear and positive association between log2-transformed serum cotinine levels and constipation risk (P for non-linear = 0.876) after adjusting all confounding factors, as shown in Figure 2.

The data are presented as a subgroup analysis in Figure 3. Subgroup analyses by age, sex, BMI, and previous diseases illustrated no statistically significant interactions (p > 0.05) in participants with cotinine levels of 0.05–2.99 ng/mL. In every subgroup, the association between smoke exposure and constipation was relatively stable.