Association of Sleep Duration and Weekend Catch-Up Sleep with Suicidal Ideation Among Adolescents with Asthma: Secondary Data Analysis Based on a Nationwide Cross-Sectional Study

Background: Sleep problems are common in asthma. We explored whether sleep duration and weekend catch-up sleep (WCUS), sleep time on weekends that exceeds sleep time on weekdays, are associated with suicidal ideation among adolescents with asthma. Methods: Data were acquired from adolescents participating in the 2019 Korean Youth Risk Behavior Web-based Surveys (KYRBS), which was self-administered, online survey. The KYRBS data were obtained from a stratied, multistage, clustered sample. Students responded prior diagnoses of asthma and sleep time. Associated factors for suicidal ideation were tested by logistic regression models. Results: Among 3,579 and 48,072 adolescents with and without asthma, respectively, adolescents with asthma were younger (14.77±1.72 vs 14.91±1.78 years old) and male predominance (57.3% vs 51.0%), had higher sleep dissatisfaction (50.1% vs. 46.2%), slept less (sleep duration 6.66 ± 1.54 vs 7.07 ± 1.51 hr), and had longer WCUS (2.94 ± 0.01 vs 2.15 ± 0.01 hr) than did those without asthma (p<0.001). The adjusted odds ratio (aOR) of suicidal ideation in those with short sleep duration (dened by ≤ 5h) was 1.41 (95% Condence Interval [CI], 1.24–1.60). Notably, long WCUS ( ≥ 2hr) was signicantly associated with a decreased risk of suicidal ideation (aOR, 0.82 [95% CI, 0.78-0.88], p<0.001). Conclusions: Although further research is needed to clarify this association, under the condition of sleep deprivation, WCUS at certain extent might be associated with low risk for suicidal ideation in adolescents with asthma.


Introduction
Although the prevalence of asthma has started to plateau and even decline in adolescents [1], asthma remains one of the most common chronic pediatric diseases with a signi cant impact on quality of life (QoL) [1]. One of the ways that asthma affects QoL is through its impact on sleep, including night asthma attacks and sleep-disordered breathing [2][3][4][5]. Nocturnal symptoms trigger frequent awakening during sleep and induce shorter sleep duration [6], which impair daily activities, increases morbidity, induces psychometric changes, reduces QoL and these changes can even contribute to suicidality [7]. However, the association between sleep duration and asthma is con icting. Some studies have found that children with asthma have generally shorter sleep durations than those without asthma [8], while other studies have not found such differences in adolescents [9]. Another report for children found that sleep duration in asthmatics is signi cantly longer than it is in non-asthmatics [10]. In adults, those with asthma were signi cantly likely to have U-shaped sleep pattern (sleep less than 6 hours or 9 or more hours as compared with those without asthma) [6]. Recently, compared with 7-8 h of sleep per night in adolescents, both short sleep duration (sleeping for < 7 h) and long sleep duration (sleeping for ≥ 9 h) had 1.22 (95% CI 1.07, 1.40) and 1.31 (1.06, 1.63) times higher odds ratio (OR) of asthma, respectively [5].
In the terms of sleep deprivation, weekend catch-up sleep (WCUS) is de ned as sleep time during the weekend that exceeds sleep time on weekdays [3,11]. WCUS is not only an indicator of insu cient weekday sleep, but also a compensatory phenomenon for weekday sleep debt [3,11]. Adolescents tend to sleep for shorter durations during the night and exhibit a delayed sleep phase than do children [12].
Adolescents also tend to sleep less than children do during school days because of the pressure of school work. In order to compensate for insu cient weekday sleep over the entire week, adolescents then tend to sleep more on the weekends [12,13]. The effects of WCUS on one's overall health are con icting. Some studies have found that WCUS is associated with a lower risk of hypertension [11], obesity prevention in adults [14], and lower frequency of asthma in adolescents [3]. In contrast, WCUS had adverse effects on executive functions, 15 and contributes to suicidal attempts and self-injury in children and adolescents [16].
Sleep disturbances in asthma including di culty in falling asleep, frequent nighttime awakenings, daytime sleepiness, and early awakenings related to its respiratory symptoms [17], are well known. Sleep is one of major modi able factors for reduction of suicide in that insu cient sleep or sleep di culties are signi cantly associated with suicidality among adolescents [18]. In the general adolescent population, a U-shape curve to sleep among adolescents in which too short or too long sleep duration is associated with mental health problems including suicidality [19][20][21]; teens sleeping ≤ 5 hours and ≥ 10 hours had more suicidal attempts than those sleeping 8 hours per night [22]. Moreover, in the point of that adolescents with asthma had greater risks of sleep disorder and suicidality than those without asthma [23], more attentions for sleep pattern in adolescents with asthma are needed for improvement of public health. Many studies exploring the association between sleep duration and asthma [5,6,[8][9][10]24], or between asthma and suicidal behaviors [23,[25][26][27], and or between sleep duration and suicidal ideation have been undertaken [19-21, 28, 29]. However, to the best of our knowledge, no prior studies have addressed the in uence that sleep duration and WCUS have on suicidal ideation in adolescents with asthma in a large population. Current suicidal ideation is considered one of the most important clinical predictors of future attempts and death by suicide [30]. Therefore, it is crucial to identify the sleep factors that are associated with suicidal ideation in adolescents with asthma in order to reduce suicide in this group.
Given this background, we hypothesized that adolescents with asthma are more likely to have shorter total and weekday sleep duration than are peers without asthma. We also suspected that adolescents with asthma would have longer weekend sleep duration and longer WCUS to compensate for their school day sleep deprivation than would adolescents without asthma. The main purpose of this study was to identify the association of sleep duration and WCUS on suicidal ideation and associated factors in adolescents with asthma after comparing the clinical characteristics, sleep parameters, and mental health status in adolescents with and without asthma in a nationally representative Korean dataset.

Study design and participants
This study was designed as a secondary data analysis using and interpreting data drawn from the third through thirteenth Korean Youth Risk Behavior Web-based Surveys (KYRBSs) completed across 7 years from 2013 to 2019. KYRBS was initiated by the Korean Ministry of Education, Ministry of Health and Welfare, and Centers for Disease Control and Prevention (KCDC) in 2005. The KYRBS is a cross-sectional, self-administered, anonymous, online survey that is conducted annually in a nationally representative sample of Korean adolescents aged 13-18 years. The understanding, reliability and validity of each question were investigated by the KCDC to qualify the surveys every year. It is known that the reliability estimates for the KYRBSs questionnaire show good validity [31]. The Institutional Review Board (IRB) of the KCDC approved this online statistical research during the entire survey every year. Survey participation excluded age-eligible respondents with absenteeism (school absence without permission), special educational needs (such as developmental disabilities), and dyslexia. In order to obtain a nationally representative sample, the KYRBS used the strati ed three stage random cluster sampling method, which included strati cation, sample allocation, and strati ed cluster sampling, as is described. The study population was strati ed by geographic region and school type to minimize sampling error. In sample allocation, 400 middle schools and 400 high schools were selected by proportional sampling to match the study population every year. In strati ed cluster sampling, sample classes were selected by simple randomization sampling from the selected schools. All of the participants provided written informed consent (both directly and from their parents or legal guardians) for approval of survey access. Every student participated voluntarily and completed the survey by logging on to the survey's Internet webpage at their school's computer laboratory using a randomly assigned unique identi cation number. The participants were not able to skip to subsequent questions without providing proper answers to every question. The online system does not accept omitted responses. Data were considered to be missing if there was any logical error or if there was outlier. In the present study, we obtained the raw data that are publicly available from the online site of KYRBSs (https://www.cdc.go.kr/yhs/home.jsp). The reasons for drop-out or non-participation were not available.

Evaluation indices 1) Asthma
Participants were asked about their history of lifelong asthma using the following question: "Have you ever been diagnosed with asthma by a doctor at any point in your life?" The accepted responses were either "yes" or "no." The disease severity, current symptoms, presence of management, and treatment modality were not assessed.

2) Demographic and socioeconomic variables
Information on demographic and socioeconomic characteristics (age, sex, school grade [middle school high school], residential type, perceived socioeconomic state [SES], and academic achievements) was evaluated. The type of residence was assessed as living with parents and without parents (such as living with relatives, living independently such as in a dormitory, or in an orphanage). The degree of SES and academic achievements were initially categorized as high, middle-high, middle, middle-low, and low.
However, we categorized these as high (high or middle-high), middle (middle), or low (middle-low or low).

3) Health-related behavioral variables
The frequency of breakfast consumption was assessed with the question, "During the past 7 days, how many days did you eat breakfast?" with allowed responses of "never, 1, 2, 3, 4, 5, 6, or 7 days." Those who indicated eating breakfast less than 2 days per week were called "breakfast skippers" according to the standards set in the survey. The frequency of vigorous exercise in the preceding 7 days, which was accompanied by very fast breathing or sweating, and by > 20 minutes of activity was also assessed.
Vigorous exercise was classi ed into none, occasional (1-3 days), and frequent (more than 4 days per week). Participants were considered to be current smokers if they answered "more than 1 day over the past month" to the question "Do you smoke?" For alcohol consumption status, the current drinker was indicated by a reply of "more than 1 day over the past month" to the question 'Do you drink alcohol?" in the standpoint of survey.

4) Emotional variables including suicidality
Psychological status was assessed using the following four variables: perceived health status, perceived level of happiness, perceived level of stress, and experience of depressive symptoms. Subjective healthiness was assessed with the following question: "What do you think about your health state?" The answer to this question was categorized as follows: healthy, average, and unhealthy. Perceived level of happiness was assessed based on a self-rating happiness scale and classi ed into happy, average, and unhappy. Perceived stress status was also assessed with the following question: "How often do you feel stress?" and was categorized as follows: often, sometimes and rarely. The following question was used to assess depression: "Within the last year, did you feel sad, blue, or depressed that caused cessation of your usual activities almost every day for two weeks or more?" This question was based on KYRBS and was answered with a binary response (yes or no). The measurement of suicidal ideation was also a binary dependent variable (yes or no) with the examining question being, "During the past 12 months, have you ever seriously thought of committing suicide?" Repetition of depressive mood and suicidal ideation were not assessed.

5) Sleep
The major independent variable for our study was sleep duration and WCUS. In the KYRBSs, self-reported wake up time and bedtime were determined on the basis of participants' responses to the following questions separately for both weekdays and the weekend: (1) "What time did you usually go to bed and wake up during the weekdays (school days) over the last week?" and (2) "What time did you usually go to bed and wake up during the weekend over the last week?" The responses were provided separately for weekdays and the weekend as follows: (1) sleep time: ( ) o'clock ( ) minute AM/PM and (2) wake up time: ( ) o'clock ( ) minute AM/PM. The responses for sleep time on both weekdays and weekends were categorized as: ≤9:00, 9:00-10:00, 10:00-11:00, 11:00 PM-12:00AM, 12:00-1:00, 1:00-2:00, and ≥ 2:00 AM. Because wake up time was in uenced by school attendance, the responses for wake up time were categorized as ≤ 5:00, 5-6:00, 6-7:00, 7:00-8:00, and ≥ 8:00 AM on weekdays, and ≤ 7:00, 7:00-8:00, 8:00-9:00, 9:00-10:00, 10:00-11:00, and ≥ 11:00 AM on weekends. We de ned the subjects who went to sleep after 2 AM as night owls, and those who woke up before 7 AM as early larks based on the previous literature [3]. Sleep duration was also separately calculated for weekdays (sleep duration on school days) and weekends (sleep duration on school-free days) on the basis of their previous responses. The participants reporting sleep less than 2 h or more than 20 h for weekdays or weekends were excluded in this study. The average sleep duration was calculated using the following weighted mean value: (5 × weekday sleep duration + 2 × weekend sleep duration) ÷ 7 [3,12], and categorized as: ≤ 5, 5-6, 6-7, 7-8, 8-9 and ≥ 9 h. Based on the recommendations from the National Sleep Foundation's sleep time duration criteria [32], the reference sleep duration was de ned by 7-8 hours per day in this study. WCUS was calculated as the average weekend sleep duration minus average weekday sleep duration [3,12], and this was divided into four categories: no catch-up, < 1 hour, 1-2 hours, and ≥ 2 hours. The four categories were used because the latter two groups occupied the rst and second largest proportions of WCUS and usually long WCUS was de ned as sleeping at least 2 h longer on weekends compared to weekdays [33].
No catch-up has the same meaning as CUS ≤ 0 hour [34]. The levels of sleep satisfaction were assessed by the degree of recovery from fatigue by sleep according to the following question: "How satis ed are you with your sleep during the last week?" We re-categorized sleep quality into the following three groups: enough (plenty and enough), a little, and not enough (not enough and never enough).

Statistical Analyses
We conducted all of the statistical analyses using complex sample procedures of the Statistical Package for the Social Sciences (SPSS) software program version 21.0 (IBM Corp., Armonk, NY, USA). Because the KYRBS data were collected through a representative, strati ed, and clustered sampling method (as a statistical representative of the general population), data from the survey were weighed based on the sample design. Descriptive statistics were used to depict the basic characteristics of the study population. The chi-square test for categorical variables and independent t-test for continuous variables were used appropriately to compare the general characteristics between subjects with and without asthma. Following the selection of signi cant covariates using univariable logistic regression analysis, multivariate logistic regression analysis was carried out to determine which sleep-related factors (e.g., sleep satisfaction, sleep duration, and WCUS) were independently contributing to the risk of suicidal ideation in adolescent with asthma: Model 1 adjusted for sex and grade; Model 2 adjusted for Model 1 variables + socioeconomic variables (residential type, SES, and academic achievement) and healthrelated behavioral variables (breakfast skipping, smoking, alcohol drinking, and vigorous exercise); and Model 3 adjusted for Model 2 variables + psychological variables (perceived status of health, perceived level of happiness, perceived frequency of stress, and experience of depressive symptoms). The results were expressed using adjusted odds ratios (OR) and 95% con dence intervals (CI). Statistical signi cance was indicated by p-values < 0.05 in all tests.

Results
Demographic characteristics of participants A total of 57,303 among 60,100 targeted adolescents completed the survey with a total response rate of 95.3%. Among them, 5,652 participants did not provide available sleep records. Therefore, their data was considered to be missing and therefore, the response rate of this study among participants was 90.1%. This left a total of 3,579 adolescents with asthma and 48,072 adolescents without asthma. The overall prevalence rate of asthma in the survey was 7.20%. Table 1 shows the general characteristics of the participants and the differences between subjects with asthma and without. Adolescents with asthma were younger (14.77 ± 1.72 vs 14.91 ± 1.78 years old) with male predominance (57.3% vs 51.0%) than those without asthma (p < 0.001). There was no statistical difference between the two groups according to SES and academic achievement. However, adolescents with asthma self-reported unhealthier and unhappier perceptions, had more stress, higher levels of depression, and more frequent suicidal ideation and suicidal attempt than did those without asthma. NOTE. Survey data were weighed for statistical representation of the general population on the basis of the sample design. The chi-square test was applied to determine statistical differences between categorical data, and the independent t-test was used for continuous variables.
Sleep parameters among adolescents with and without asthma Table 2 shows the differences of sleep parameters between the asthma groups and non-asthma groups. Adolescents with asthma were signi cantly different from those without asthma in terms of every sleep variable. Most of all, sleep dissatisfaction was signi cantly higher (sleep satisfaction was not enough: 50.1% vs. 46.2%) in the asthma group than it was in the non-asthma group. Adolescents with asthma slept less than did those without asthma (average sleep duration 6.66 ± 1.54 vs 7.07 ± 1.51 hr, proportion of less than 5 hr of sleep duration was 10.9% vs 8.4%, respectively, p < 0.001). Asthmatics were also more likely to be night owls on both weekdays (31.4% vs 27.4%) and weekends (49.2% vs 48.0%), and to be early larks on weekdays (34.0% vs 31.4%), while less likely to be on weekends (3.6% vs 5.5%). The weekday sleep duration (5.72 ± 1.51 vs 6.48 ± 1.48 hr) was shorter, while the weekend sleep duration (9.01 ± 2.89 vs 8.53 ± 2.83 hr) was signi cantly longer in adolescents with asthma than it was in those without. In addition, the WCUS duration was signi cantly longer in the adolescents with asthma than it was in those without asthma (2.94 ± 0.01 vs 2.15 ± 0.01 hr). The proportion of longer WCUS among adolescents with asthma was 43.8%, and that among adolescents without asthma was 43.1%.   Abbreviations: OR, odds ratio; CI, con dence interval. NOTE. After the selection of signi cant covariates, univariate and multivariate logistic regression analyses were performed to identify the associated factors for suicidal ideation in adolescents with asthma. Model 1 adjusted for sex and grade; Model 2 adjusted for Model 1 variables + socioeconomic variables (residential type, socioeconomic state, and academic achievement) and health-related behavioral variables (breakfast skipping, smoking, alcohol drinking, and vigorous exercise); and Model 3 adjusted for Model 2 variables + psychological variables (perceived status of health, perceived level of happiness, perceived frequency of stress, and experience of depressive symptoms). Weekend catch-up sleep was calculated as the average weekend sleep duration minus average weekday sleep duration, and this was divided into four categories: ≤0 hour, 0-1 hour, 1-2 hours, and ≥ 2 hours. No weekend catch-up sleep has the same meaning as catch-up sleep ≤ 0 hour.

Discussion
The main nding of our study was that sleep dissatisfaction, short sleep, and asthma were all associated with an increased risk of suicidal ideation. However, long WCUS was associated with a decreased risk of suicidal ideation in Korean adolescents with asthma. To the best of our knowledge, this is the rst study to investigate the relationship between WCUS and suicidal ideation in adolescents with asthma after adjusting for multiple potential confounding factors.
Our results indicate that short sleep duration is signi cantly associated with suicidal ideation in adolescents with asthma. Approximately 8-10 hours and 7-9 hours of sleep are considered su cient for teenagers and young adults, respectively [32]. In general, adolescents are considered to be chronically sleep deprived with decreased total sleep duration and increased daytime sleepiness [9]. This chronic sleep deprivation is due to a variety of factors including decreased parental control, inconsistent bedtimes, increased schoolwork, poor sleep hygiene (e.g., caffeine use late in the day, technology in the bedroom, smartphone overuse), and a discordance between biological circadian rhythms and early school start times [9]. In Korea, most middle-and high-school classes start before 8:30 AM. And many Korean adolescents attend private academic institutions and study late at night even on the weekends [12]. These students face strong psychosocial pressure to attain su cient grades for admission into prestigious universities [5,12]. The link between short sleep duration and suicidal ideation in adolescents with asthma is complex and not fully understood. However, there are several possible explanations for this association.
First, sleep deprivation itself is considered to be a stress-inducing factor for adolescents with a higher risk of several mental disorders such as behavioral disorders [28], substance abuse [28], depression [28,29], and suicidality [28]. In adolescents with asthma who have sleep problems commonly, including prolonged sleep onset latency, di culty in falling asleep, frequent nocturnal awakenings, daytime sleepiness, and early awakenings [17], even when their asthma symptoms are well controlled [9], short sleep duration may be associated with higher risk of suicidal ideation In addition, given its long-term disease course, asthma is associated with depression due to the need for chronic treatment and the repeated exacerbations that may themselves cause hopelessness [25]. The negative in uence of asthma on mental-health states is widely accepted; people with asthma had a signi cantly increased risk of depression than did controls [23,27]. Adolescents with asthma may potentially experience more psychological burden from the disease than do adult patients [23,35,36].
Longitudinal experience with asthma can be associated with more debilitating respiratory symptoms of wheezing, cough, shortness of breath, and near-death episodes, which could certainly generate fearful beliefs about respiratory symptoms, which might in turn provoke separation anxiety, panic attacks, and depression [23,27]. The chronic nature of this disease may lead to increased asthma symptoms through risk-taking behaviors (smoking, drinking, non-adherence to medication and poor illness control in general) 25 and consequently increased suicide risk [26,27]. As it is known that people with asthma have a signi cantly increased risk of suicidal ideation [23], suicidal ideation may become more prevalent in adolescents with asthma when they become sleep deprived.
Third, in patients with asthma, the prevalence of obstructive sleep apnea (OSA) and sleep disordered breathing (SBD) is consistently higher than it is in the population without asthma [17,37]. Asthma seems to be a consistent and independent risk factor for SDB (OR = 1.9 [1.69-2.18]) [37]. Gutierrez et al. found that maximal percentage of rapid eye movement (REM) SpO 2 desaturation, and the REM obstructive apnea-hypopnea index (OAHI) were signi cantly higher in asthmatic children with OSA than they were in subjects with OSA alone [38]. These ndings may result in sleep deprivation in adolescents with asthma, and then may be associated with a high risk of suicidal ideation related to the aforementioned negative mental health state in asthma.
Another notable nding in our study was that adolescents with asthma were more likely to be night owls on both weekdays and weekends, while be early larks only on weekdays. As we hypothesized, adolescents with asthma had shorter weekday sleep duration and longer weekend sleep duration than did those without asthma. These ndings may be explained by the sleep characteristics and problems of asthma itself, including nocturnal symptoms and the aforementioned educational uniqueness in Korea.
These differences may also be related to the longer WCUS in adolescents with asthma than that of those without asthma for sleep debt compensation. Because individual sleep needs vary, WCUS is not only an indicator of the degree of insu cient weekday sleep [3], but also is a compensatory behavior to cope with weekday sleep debt [33]. A recent study found that the risk for anxiety and depression was higher among adults who did not have WCUS than it was among those who did [33]. Long WCUS was associated with a decreased frequency of asthma and depression in Korean adolescents [3,12]. However, it is unclear whether WCUS has advantageous effects on emotional states of adolescents with asthma, because sleep hygiene guidelines recommend regularizing sleep time and wake-up time [12,32]. In particular, the association between suicidal ideation and WCUS in adolescents with asthma has not been evaluated previously. To the best of our knowledge, our results are the rst to show that su cient WCUS duration (of ≥ 2hr) may play a protective role against suicidal ideation among adolescents with asthma. However, this nding does not mean that adolescents with asthma should sleep longer on the weekends than they do during weekdays. Nevertheless, it is important for adolescents with asthma to achieve su cient sleep time during the weekend in the way that WCUS is an indicator of insu cient weekday sleep.
There are several methodological limitations in this study that must be noted. First, our results are subject to recall bias, because we used self-reported data collected via the internet. Studies using more objective assessments of sleep, such as polysomnography or actigraphy, are needed to avoid this limitation.
Second, we only collected data at a single point in time by asking about typical sleep pattern over the preceding week. Weekend sleep duration may vary depending on the situation of the following week (e.g. school examination). Longitudinal studies are needed to clarify our ndings, as asthma control and sleep patterns may change throughout the school year. Third, this study was cross-sectional in design; therefore, we were unable to establish a causal relationship between associated sleep factors and suicidal ideation in adolescents with asthma. Well-controlled prospective studies are necessary to con rm the results of the present study. Fourth, we did not analyze asthma management, treatment modalities, or asthma severity because such information was lack and only the lifelong diagnosis of asthma was evaluated in the survey. This omission may have led to the misclassi cation of asthma. In addition, despite adjusting for various potential confounders, the survey did not assess other signi cant factors that have been associated with poor sleep quality, such as OSA, snoring, nocturnal gastroesophageal re ux, internet or smartphone use time, and caffeine or drug use.
Despite these limitations, this study also has several strengths and important clinical implications for adolescents with both asthma and sleep problems. This was a nationwide, government-directed survey with a high response rate (96.2%). To the best of our knowledge, our study included the largest number of participants among similar studies. A socioeconomically diverse sample with an equal proportion of middle school and high school students (400 schools each annually) were represented. All of the analyses in this study were based on sample weights and were adjusted for the complex sample design of the survey. Almost all participants had the same ethnic background, which minimizes other possible confounding factors. These study characteristics also mean that our results are generalizable to the overall Korean population. Most of all, this research was the rst to identify associated sleep factors associated with suicidal ideation in adolescents with asthma. No prior study had evaluated the relationship between WCUS and asthma. The public health implications of our ndings can be applied at the educational and practical levels. A recent study demonstrated that adolescents were able to successfully extend sleep duration with sleep hygiene education [39]. This simple intervention would likely be bene cial for adolescents with asthma [9]. From an educational perspective, strategic interventions to reduce academic stress on Korean adolescents should be seriously considered, such as later start to classes. Health care professionals, including pediatric allergists, require a greater awareness of the risk of mental health problems and suicidality in adolescents with asthma. This is particularly true in those with concurrent sleep problems. When treating adolescents with asthma, these patients require a holistic approach with a more accurate assessment of associated sleep factors related to suicidality.
Our study protocol for secondary analysis and interpretation was reviewed and approved by the IRB of Samsung Changwon Hospital (IRB No. 2020-06-009). Informed consent was exempted by the board, because our study was based on the secondary data analysis.

Funding
No external funding sources reported.