U-shaped association between sleep duration with chronic constipation and diarrhea: A population-based study

ABSTRACT To examine the association between sleep duration with chronic constipation and diarrhea, we collected demographic and questionnaire data of participants from The National Health and Nutrition Examination Survey in the period 2005 to 2010. A restricted cubic spline curve function was used to determine the association between sleep duration and chronic constipation or diarrhea. Multivariate logistic regression models were used to estimate the association between sleep duration groups and chronic constipation or diarrhea. 14,054 eligible participants were included in this study. Among all sleep durations, 7 hours sleep per day had the highest percentage of normal stool forms (86.3%, P < .001), while people with ≤4 hours sleep had a higher number of bowel movements per week (P < .001). RCS models demonstrated a significant nonlinear association between sleep duration and risk of chronic constipation (P < .001). The inflection points of the U-shaped association curve corresponded to 7 hours of sleep per day. Multivariate logistic regression indicated that compared to participants with 7 hours daily sleep duration, participants with ≤4 hours and ≥10 hours sleep duration had a 54% (95%CI: (1.16–2.04), P = .002) and 90% (95%CI: (1.33–2.72), P < .001) higher risk of constipation, respectively. There is a non-linear association between sleep duration and the risk of chronic constipation. Our findings indicate that both shorter and longer sleep duration are associated with an increased risk of constipation.


Introduction
Chronic constipation and chronic diarrhea are common gastrointestinal diseases in the general population, with an estimated prevalence of 6.1% -28.0% in Asia (Ono et al. 2018;Zhao et al. 2012). The current Rome criteria classify this group of diseases as functional gastrointestinal disorders (FGID) characterized by persistent and recurrent gastrointestinal symptoms (Barberio et al. 2021;Drossman 2016;Lacy and Patel 2017). In recent years, numerous studies have indicated that functional gastrointestinal disorders can have an effect on patients' sleep and cause sleep disorders (Futagami et al. 2016;Jansen et al. 2021;Wuestenberghs et al. 2022;Zhao et al. 2018).
Sleep is the major physiological process that regulates circadian rhythms for humans (Zimmet et al. 2019). When the body's circadian rhythm is not synchronized with the sleep schedule, it may lead to circadian rhythm misalignment (Archer and Oster 2015). Several studies have indicated that the sleep-related light-dark cycle and circadian rhythm regulation have a direct effect on the gastrointestinal tract Ghabril et al. 2017). Large epidemiological studies have proved that sleep insufficiency and mistimed sleep contribute to obesity, type-2 diabetes, inflammatory bowel disease (IBD), and other diseases (Archer and Oster 2015). Furthermore, clinical studies have indicated that patients with IBD often have subjective sleep disorders (Vgontzas et al. 2004).
Many studies have indicated that patients with chronic constipation or diarrhea have significantly lower mean physical and mental component scores than controls (Brochard et al. 2019;Sun et al. 2011). And FGID can lead to impairment of health-related quality of life, such as IBD and musculoskeletal disorders (Belsey et al. 2010). However, few studies have analyzed the association between sleep duration, an indicator associated with a healthy lifestyle, with FGID. Therefore, we performed a study on the association between sleep duration with chronic constipation and diarrhea, based on data from a large population-based survey program.

Data source
The National Health and Nutrition Examination Survey (NHANES) is a cross-sectional program conducted by the National Center for Health Statistics (NCHS), to monitor the physical and nutritional status of the U.S. population. The program uses interviews, professional medical examinations, and laboratory tests to obtain data on participants' demographic information, dietary status, and personal medical care data. The National Health Institutional Review Board (NHIRB) has approved this project and all participants or guardians of participants have signed an informed consent form ).
The samples with missing bowel health and sleep questionnaires were excluded. Samples combined with specific medications (n = 751, Including laxatives and other drugs that may impact gastrointestinal function), pregnancy (n = 412) and gastrointestinal malignancy (n = 133) were also excluded ( Figure 1).

Bowel health questionnaire
The bowel health questionnaire in NHANES was used to identify participants suffering from chronic constipation and chronic diarrhea. Participants were shown a card with colorful pictures and descriptions of the seven Bristol Stool Form Scale (BSFS) types and were asked to "Please look at this card and tell me the number that corresponds with your usual or most common stool type for the past 12 months" . Participants who identified their usual or most common stool type as BSFS 1 (separate hard lumps, such as nuts) or BSFS type 2 (sausage-like, but lumpy) were classified as chronic constipation. Those participants who took laxatives once a week or more in the past 30 days were also classified as having chronic constipation. Participants who considered their usual or most common type of stool to be BSFS type 6 (fluffy-edged stool with paste-like stool) or BSFS type 7 (watery, no solid masses) were classified as having chronic diarrhea. The rest of the participants were considered to have normal bowel movements.

Sleep information
In the sleep questionnaire information, we collected the self-reported daily sleep duration of participants in the past 12 months. We divided them into seven groups according to daily sleep duration, ≤4 hours, 5 hours, 6 hours, 7 hours, 8 hours, 9 hours, and ≥10 hours. We merged ≤4 and ≥10 hours of sleep per day due to their small number. The sleep disorder diagnoses in this study were obtained from physicians or other health professionals and included sleep apnea, insomnia, restless legs, and other related etiologies.

Covariates
To minimize confounding factors that may have an effect on sleep, we adjusted age (<20, 20-60, >60), gender (male, female), ethnicity (non-Hispanic white, non-Hispanic black, Mexican American, Hispanic, and other races), education level (below high school, high school or equivalent, college or above), marital status (married and non-married), poverty income ratio (PIR) (≤1.30, 1.31-3.5, >3.5), body mass index (BMI) (<25 kg/m 2 and ≥25 kg/m 2 ), recreational activities, military service history, drug use (include marijuana, heroin, methamphetamine, etc.), smoking status and alcohol abuse (female ≥10 grams or 1 cup/day, male ≥20 grams or 2 cups/day) in the study. Complications (including asthma, tinnitus, hypertension, diabetes, hyperlipidemia, congestive heart failure (CHF), renal failure, coronary heart disease (CHD), and stroke) were also included in the study. The diagnosis of complications either met the clinical criteria or relied on the medical history in the questionnaires. We also performed univariate analysis of all covariates to estimate their effect on chronic constipation or diarrhea (Table S1).

Statistical analyses
Continuous variables were expressed as the median and interquartile range (IQR), and categorical variables were expressed as n (%). For categorical variables, P values were analyzed by the χ2 test. For continuous variables, the Mann-Whitney U test was used in the non-normal model.
Restricted cubic spline (RCS) was used to fit the nonlinear relationship between variable X (risk of chronic constipation or chronic diarrhea) and variable Y (sleep duration). As recommended by Harrell et al, we set the nodes in the RCS model to 5 and corrected for the covariates (Herndon and Harrell 1995). The association between sleep duration with chronic constipation or chronic diarrhea was assessed using univariate and multivariate logistic regression methods, and adjusted odds ratios (AOR) and 95% confidence intervals (CI) were calculated. In the multivariate logistic regression analysis, we constructed three models A, B, and C to correct for the relationship between sleep duration and stool form. Model A included 6 variables from the demographic information; Model B added 10 comorbidity variables to Model A; And model C added variables such as BMI, smoking status, alcohol abuse, drug use, military service, and recreational activities to Model B.
Statistical analyses were performed using SPSS software (version 24.0), R (4.0.5), and Rstudio software (version 1.4.1106). Probability values of P < .05 were considered to be statistically significant.
After correction by combining the weights of the three cycles as recommended by NHANES, 7 hours sleep duration per day had the lowest rate of sleep disorders with only 5.5% (Table 1). We also observed that 7 hours sleep duration per day had the highest percentage of normal stool form, while those with ≤4 hours sleep duration had a significantly higher number of bowel movements per week. There were 298, 751, 3463, 7324, 1123, 952 and 143 participants in the seven different BSFS types (Type 1-7). Sleep duration was significantly lower in the sleep disorder group than in the normal sleep group, and this discrepancy was observed in almost all stool types of participants ( Figure 2A). Furthermore, we found that among the seven different BSFS types, chronic constipation (Type 1, 9.7%) and chronic diarrhea (Type 6 and 7, 10.4% and 15.4%) had significantly more patients with sleep disorders than the normal bowel movements group (Type 3, 4, 5) ( Figure 2B).
By using the RCS model we found a U-shaped association between sleep duration and the risk of chronic constipation (P < .001 for nonlinearity, Figure 3). It can be found that the risk ratio of chronic constipation tends to increase whether the sleep time increases or decreases with a threshold value of 7 hours sleep duration ( Figure 3A). A similar trend can be observed in the relationship between chronic diarrhea and sleep duration, despite its nonlinear p-value >0.05( Figure 3B).

Discussion
In this study based on large cross-sectional data of the general United States population, we found an association between sleep disorders and functional gastrointestinal disorders and a nonlinear U-shaped association between sleep duration and chronic constipation, with 7 hours of sleep per day corresponding to the lowest risk of chronic constipation. Sleep durations of <7 hours and >7 hours per day were both associated with a high risk of chronic constipation.
So far, there were few reports on the association between FGID and sleep duration, and due to the low level of scientific consideration, most perspectives suggest that FGID reduces the quality of sleep and cause sleep disorders in patients (Brochard et al. 2019;Liu et al. 2021;Zhang et al. 2021). Results from the United States questionnaire and a large Japanese internet survey both showed a significant correlation between sleep Relative risk for chronic constipation and diarrhea according to sleep duration in Logistic Regression model. All models were adjusted for covariates such as age, gender, ethnicity, marital status, education, PIR, comorbidity, BMI, smoking status, alcohol and drug abuse, military service, recreational activity, and sleep disorder.
scores and gastrointestinal symptom scores, patients with FGID had significantly less sleep compared with control subjects (Cremonini et al. 2009;Yamamoto et al. 2021). In this study, we also found that the proportion of sleep disturbances was indeed higher in the chronic constipation (BSFS 1, 2) and diarrhea (BSFS 6, 7) population than in the normal stool group (BSFS 3,4,5). However, the reality of the association does not seem to be such a monotonic effect.
We can observe the association between sleep duration and chronic constipation or diarrhea fitted using the RCS model which has a U-shaped curve with 7 hours of sleep duration per day as the inflection point. Sleep duration more or less than 7 hours corresponds to an increased risk of chronic constipation or diarrhea. This association remained stable between chronic constipation and sleep duration after bringing the results into a logistic regression model for multi-covariate adjustment. This seemingly contradictory conclusion can be supported by studies in recent years. Large-scale population data from the Asia Cohort Consortium and UK Biobank both demonstrated that self-reported sleep duration of 7 hours per day corresponds to the lowest risk of all-cause mortality or cognitive decline (Li et al. 2022;Svensson et al. 2021).
However, no scientifically valid explanation can be found for the observation that prolonged sleep duration corresponds to an increased risk of FGID. The commonly held view at this stage supports that the association between FGID and sleep disorders is bidirectional (Ananthakrishnan et al. 2014;Stevens et al. 2017;   Wilson et al. 2015). FGID activity may affect rapid eye movement sleep through elevated circulating cytokines, resulting in sleep disorders (Nojkov et al. 2010;Vgontzas et al. 2004). And the accumulated emotional stress of sleep disorders in turn activates the hypothalamicpituitary-adrenal (HPA) axis, which in turn affects the composition of the gut microbiota, exacerbating gastrointestinal inflammation (Bailey et al. 2011;Cryan and Dinan 2012;O'Mahony et al. 2009). Although the causal association is not clear, we can still infer that treating either of them, whether constipated patients complain of sleep disorders or sleep-disordered patients complain of concomitant constipation, will bring about improvements in the patients' symptoms.
Of course, there are some limitations to our study. First, the data on sleep duration was based on participants' selfreports rather than recorded by polysomnography, which may cause some bias. And sleep questionnaires lack objective measures such as sleep quality. Second, many factors may affect sleep or stool form under realistic conditions (including lifestyle, dietary habits, and special prescriptions). These factors were not included in this study, which may also introduce bias. Last but not least, crosssectional studies cannot show a causality relationship between sleep duration and chronic constipation, which requires further longitudinal studies or animal research.
In conclusion, our study data suggest a U-shaped association curve between sleep duration and the risk of chronic constipation. A daily sleep duration of 7 hours may be optimal.