|Home | About | Journals | Submit | Contact Us | Français|
To explore the relationship between sleep-disordered breathing (SDB) and behavioral problems among inner-city children with asthma.
We examined data for 194 children (age 4–10 yrs) who were enrolled in a school-based asthma intervention program (response rate: 72%). SDB was assessed using the Sleep-Related Breathing Disorder (SRBD) questionnaire that contains 3 subscales: snoring, sleepiness, and attention/hyperactivity. For the current study, we modified the SRBD by removing the 6 attention/hyperactivity items. A sleep score of >.33 was considered indicative of SDB. To assess behavior, caregivers completed the Behavior Problem Index (BPI) which includes 8 behavioral subdomains. We conducted bivariate analyses and multiple linear regression to determine the association of SDB with BPI scores.
The majority of children (mean age 8.2 yrs) were male (56%), African American (66%), and insured by Medicaid (73%). Overall, 33% of children experienced SDB. In bivariate analyses, children with SDB had significantly higher (worse) behavior scores compared to children without SDB on total BPI (13.7 vs 8.8, p<.001), externalizing (9.4 vs 6.3, p<.001), internalizing (4.4 vs 2.5, p<.001), anxious/depressed (2.4 vs 1.3, p<.001), headstrong (3.2 vs 2.1, p<.001), antisocial (2.3 vs 1.7, p=.013), hyperactive (3.0 vs 1.8, p<.001), peer conflict (.74 vs .43, p=.011), and immature (2.0 vs 1.5, p=.001). In multiple regression models adjusting for several important covariates, SDB remained significantly associated with total BPI, externalizing, internalizing, anxious/depressed, headstrong, and hyperactive behaviors. Results were consistent across SBD subscales (snoring, sleepiness).
We found that poor sleep was independently associated with behavior problems in a large proportion of urban children with asthma. Systematic screening for SDB in this high-risk population might help to identify children who would benefit from further intervention.
Asthma is one of the leading causes of childhood illness,1–3 affecting nearly 9 million children in the United States.4 The public health burden of childhood asthma is extensive including high rates of hospitalizations2,5–7 and emergency department visits,3, 5,8,9 absenteeism from school and work,5,10 and impaired quality of life.11,12 Furthermore, children from poor and African American backgrounds suffer disproportionately from asthma.8,13,14
Several studies have linked asthma symptoms and childhood behavior problems, such as hyperactivity and inattention.15–17 For example, a meta-analysis of 26 studies found that children with persistent asthma symptoms had higher levels of behavioral problems compared to healthy children.16 Troubled behavior among children with asthma may be compounded by sleep-disordered breathing (SDB) which encompasses a continuum of sleep related disturbances ranging from primary snoring to overt obstructive sleep apnea (OSA). SDB is becoming increasingly recognized in children, with prevalence estimates ranging from 0.7%–3% for the more severe OSA18–21 to as high as 7–25% for the milder form of SDB, primary snoring.19,22–25 SDB in childhood may also be associated with persistent wheezing or asthma.20, 26,27
There has been growing evidence that children with SDB have higher rates of behavioral problems compared to children without SDB 24,25,28,29, and improvement in sleep-disordered breathing may alleviate behavioral problems in children.30,31 However, limitations in the published literature preclude us from drawing a causal link between sleep and behavior,32 and additional studies are needed to help develop a more complete understanding of this relationship. Research in high-risk populations may be particularly useful since one would expect associations may be stronger in more susceptible groups of individuals.
We are not aware of studies that have specifically explored the relationship between SDB and behavior in a community sample of non-referred urban children with asthma. It is important to assess the relationship between SDB and behavior among this sample, as this is a high risk group of children for both sleep and behavior problems that may particularly benefit from appropriate interventions. In this study, we explored the association between SDB and troubled behaviors in a sample of urban children with asthma. We hypothesized that children with parent-reported poor sleep will have worse behaviors.
This study utilizes data collected from an ongoing school-based asthma intervention.33 This randomized controlled trial is designed to evaluate the impact of school nurse-administered maintenance asthma medications and an environmental tobacco smoke-reduction program for inner-city children in Rochester, NY. Our analysis includes a community sample of 226 children, ages 3–10 years, enrolled in the program (overall response rate=72%).
For enrollment, we identified children through school health forms, and a screening form was administered by telephone with the child’s primary caregiver to determine eligibility for the intervention. Children with physician-diagnosed asthma and persistent symptoms in the past year based on national guidelines34 were eligible. Written informed consent was obtained from all primary caregivers and assent was obtained from children >7 years of age prior to enrollment in this study.
From August 2006 to November 2006, each participating family received an extensive home visit to collect baseline data, including demographic information, asthma symptom severity, medications, healthcare utilization, child behaviors and caregiver factors. Families received a follow-up telephone call each month to discuss the child’s asthma symptoms and healthcare utilization. At the end of the school year (approx. June 2007), we conducted an extensive final follow-up telephone call. We assessed symptoms, child behavior, sleep problems and additional information during this final interview. We also collected a saliva sample for cotinine measurement and measured the child’s height and weight at the end of the school year.
For the current analysis we excluded 16 children without final follow-up data (7 withdrawn, 8 lost to follow-up, 1 incomplete data) and 2 children less than 4 years old because the behavior scale used here is not validated for this age group. We also excluded 14 children with an Autism diagnosis (including Autism, Asperger Syndrome, and Pervasive Developmental Disorder). Our final analytic sample included 194 children. The University of Rochester’s Institutional Review Board approved the study protocol.
We assessed childhood behavior using the previously validated Behavior Problem Index (BPI).35 The BPI was created by Peterson and Zill using many of the same questions as Achenbach’s Childhood Behavior Checklist.36 This 32-item survey is used to assess behaviors during the previous 3 months for children 4–17 years of age; 28 items are included in the survey for children 12 years and younger. Caregivers are asked to respond to statements of behavior by reporting whether each behavior is “Not true”, “Sometimes true”, or “Often true” of their child. All positive responses (sometimes true and often true) are scored as a 1 and summed to create a total behavior score (range 0–28). A score of >14 indicates significant behavior problems.37 The BPI can also be divided into several subscales; externalizing (18 items), internalizing (10 items), anxious/depressed (5 items), antisocial (6 items), hyperactive (5 items), headstrong (5 items), peer conflict (3 items), and immature (4 items). Examples of statements included in the BPI include, “Has trouble getting along with others,” “Demands a lot of attention,” and “Is too fearful or anxious.”
We assessed sleep problems by using the 22-item Sleep-Related Breathing Disorder (SRBD) scale.38,39 The SRBD is a validated subscale of the Pediatric Sleep Questionnaire (PSQ),38 and contains questions about snoring, sleepiness, apnea, attention and hyperactivity. Parents respond to questions about sleep-related behaviors by responding, “Yes” =1, “No” =0 or “Don’t know” = missing. The mean response from non-missing items creates a score between 0 and 1. A sleep score >.33 has been effective in identifying pediatric sleep-disordered breathing using polysomnygraphy as the gold standard assessment.38,39 Use of the SRBD scale in the current study was permitted by creator Ronald D. Chervin, MD, MS (personal communication, 2007).
The SRBD contains 3 subscales: snoring, sleepiness, and attention/hyperactivity. Because we evaluated the relationship between SDB and behaviors, including hyperactivity, we modified the SRBD to exclude the six attention/hyperactivity questions. These items include statements about difficulty organizing tasks, fidgeting, and interrupting conversations. Analysis of the SRBD without the 6 attention and hyperactivity questions is consistent with prior work of Dr. Chervin.40
We examined child, caregiver, and environmental covariates for this analysis. Child factors in this study consist of standard demographic variables for each child including, sex, race (white/African American/other), ethnicity (Hispanic/not Hispanic), and child’s age. We also included Medicaid insurance (yes/no), prematurity (yes/no), body mass index (age and gender adjusted Z-score), and current asthma severity (intermittent/persistent) as other variables that may be related to sleep and behavior. We assessed asthma severity during the final follow-up interview by asking parents to report the number of days in the previous 14 days their child had daytime asthma symptoms and the number of days with nighttime asthma symptoms. Children with ≥5 days of daytime symptoms or ≥2 nights with asthma symptoms during the past two weeks were considered to have persistent asthma symptoms based on national guidelines.34
Caregiver factors include the caregiver’s age (<30/≥30 years), caregiver’s education (<high school/≥high school), parent depression, parent stress, and parent quality of life.
We evaluated parent depression using the Kessler Psychological Distress Scale (K10).41 The K10 is a 10-item scale used to assess symptoms of depression and anxiety. We asked caregivers how frequently they experienced each item (e.g., nervous, depressed) in the past 4 weeks (“None of the time” (score=1) to “All of the time” (score=5)). We summed scores from all items and higher scores indicate a higher risk of depression, anxiety or both (range 10–50). We then divided scores into four categories (Well, Mild, Moderate, and Severe psychological distress) based on previously validated domains.42 We measured parent stress using questions from the competence subscale of the Parenting Stress Index with permission from the publisher, Psychological Assessment Resources, Inc.43 We included 5 items on a 5-point scale, and summed scores for a total parent stress score (range 5–25). Higher scores indicate increased parental stress.
Caregivers also rated their quality of life using the Pediatric Asthma Caregivers Quality of Life Questionnaire (PACQLQ) by Juniper et al.44 Parents answered 13 questions about how their child’s asthma may have interfered with normal daily activity over the past week. PACQLQ questions are rated on a 7-point likert scale with 1 being “All of the time” and 7 being “None of the time.” Responses were averaged for a mean quality of life score (range 1–7).
Environmental tobacco smoke exposure was measured by both parent report and the child’s level of salivary cotinine. Cotinine, a metabolite of nicotine, is used as a biomarker for many intervention studies for young children with asthma.45,46 We collected saliva samples from each child during the time of the final follow-up assessment using standard collection techniques. All samples were measured with a standard enzyme-linked immunosorbent assay and reported in nanograms per milliliter (ng/mL).
Analysis was performed using SPSS version 15.0 software (Statistical Product and Service Solutions 15.0; SPSS Inc, Chicago, Ill). We conducted student t-tests to compare mean BPI scores for children with (sleep score >.33) and without (sleep score ≤.33) sleep-disordered breathing. Multiple linear regression analyses were conducted to determine whether children with SDB have more behavior problems. Initial covariates included in the regression analysis included demographic variables, covariates that were significant in the bivariate analysis, and key exposure variables (caregiver smokes, salivary cotinine, asthma severity, preventive asthma medication, and treatment group). We performed backward stepwise regression to include covariates with selection criteria of p<.20 for entry and p<.15 to remain in the model. These analyses were repeated for the behavioral subscales of the BPI. A 2-sided alpha <.05 for the primary hypothesis was considered statistically significant. With our sample size, we estimated that we could detect a three-point difference in total BPI scores between children with and without SDB, with 80% power and an alpha value of .05.
Overall, the majority of children was male (56%), African American (66%), and insured by Medicaid (73%). Some children were born prematurely (11%), and the average age of the children was 8.2 years. Twenty-six percent of children had persistent asthma symptoms during the previous two weeks, and most children were prescribed a preventive asthma medication (86%). One third of parents had less than a high school education, and 41% of children lived with at least one smoker in the home (Table 1).
In this sample, children’s sleep scores ranged from 0 to .88 with a mean score of .27 (SD = .20). Overall, 33% of the children had a sleep score >.33, indicating SDB. Children with SDB were more likely to be female, have a parent with less than a high school education, have nighttime asthma symptoms, and have a higher body mass index (Table 1). In addition, the parents of children with SDB were significantly more stressed, depressed, and had a lower quality of life. There were no differences in age, race, ethnicity, insurance status, exposure to tobacco smoke, prematurity, asthma severity, use of preventive asthma medication or treatment group between children with and without SDB.
Table 2 shows the scores on the total BPI and the behavior subscales. Overall, 32% of children had a total BPI above 14, indicating a significant behavioral concern that may warrant professional intervention. Compared to children with no sleep difficulties, children with SDB scored significantly higher on the total BPI (13.84 vs 8.9, p<.001) and on each of the behavior subscales (all p<.05).
The results of multiple linear regression analyses are shown in Table 3 using the sleep score as a dichotomous variable (SDB/no SDB). The initial model for each regression contained the following covariates: child’s age, race, ethnicity, gender, parent education, BMI Z-score, parent smoking status, cotinine level, parent depression, parent stress, parent quality of life, asthma severity, nighttime asthma symptoms, use of preventive asthma medication, and treatment group. Using the backwards stepwise regression, many of these covariates did not remain in the final models which are indicated in Table 3. Overall, SDB remained significantly associated with total BPI, externalizing, internalizing, anxious/depressed, headstrong, and hyperactive behaviors when controlling for pertinent covariates (Table 3).
Because nighttime symptoms of asthma could be confused with sleep disordered breathing symptoms, particularly for items on the sleepiness subscale, we repeated the analyses using scores on the sleep subscales (snoring and sleepiness) separately. In each of these analyses we found similar significant associations shown between higher sleep scores and worse behaviors (results not shown).
This study examined the association between sleep-disordered breathing and childhood behavior problems among a group of inner-city children with asthma. We found that one third of urban children with asthma may be suffering from SDB. This is considerably higher than the current estimates of SDB in children, and suggests that routine screening for SDB might be particularly important for children with asthma. Similarly, we found that 32% of these children have behavioral symptoms severe enough to warrant further evaluation. It is clear from these findings that urban children with asthma are at risk for both SDB and poor behavior.
Children with asthma and sleep-disordered breathing had worse behavior compared to children without sleep difficulties. Children with SDB scored nearly 5 points higher on the total BPI compared to children without SDB. This means that children with SDB exhibited, on average, 5 more problem behaviors than children without sleep difficulties. These findings were particularly prominent in the externalizing domains, including hyperactivity and headstrong behaviors, and remained even when controlling for important variables which can influence children’s sleep, behavior, and parent’s report of sleep and behavior.
Prior research has also demonstrated a relationship between sleep disturbances and externalizing behavior problems. For example, a recently published study of children with a clinical diagnosis of SDB compared children’s scores on the SRBD scale (excluding the 6 attention/hyperactivity questions) with their behaviors before and after adenotonsillectomy.30 The authors reported a strong association between children’s sleep scores and inattention, oppositional behaviors, and an attention deficit and hyperactivity disorder index prior to surgery, with some suggestion of improvement after surgery. Another study found similar results among a large population based sample of children, where symptoms of SDB were present in 25% of children,25 and these children were more likely to exhibit problem behaviors such as hyperactivity, inattention and aggressive behaviors.
A strength of our study is that we were able to account for many possible confounding variables in our analyses, which is noteworthy since the etiology of poor childhood sleep is complex. For example, African American children, children that are overweight or obese, and children who were born prematurely have been found to be at an increased risk for SDB.20,47,48 Furthermore, mental distress of parents has also been associated with poor child sleep.49,50 Similarly, behavior problems are more prevalent among poor and urban populations.51 Social and environmental stressors such as socioeconomic status, familial issues, and environment can influence both asthma and behavioral outcomes.52
Our study is unique in that it explores the relationship between sleep and behavior among a non-referred, community-based sample of children, using previously validated surveys. Much of the literature assessing sleep and behavior has included children from clinical practices for sleep assessment, behavioral assessment or adenotonsillectomy.32 Our study observes the association between sleep and behavior among a group of non-referred, urban children with asthma, a population that could potentially benefit substantially from assessment and intervention.
Lastly, BPI and SDB surveys used in our study are brief and could be utilized in a clinical setting to help identify children with poor sleep or troubled behavior. Several studies have utilized the BPI as a parent-report measure to assess behavioral problems in children.15, 53, 54 In addition, the use of the SRBD scale in research as a way to identify children with sleep-disordered breathing is increasingly common.30, 39, 55
There are some limitations to this study. First, this is a cross-sectional study and therefore, we cannot establish a directional relationship between SDB and behavior problems in this sample of children. Additionally, behavior problems were assessed by caregiver only and were not confirmed with physicians, teachers, or with subsequent assessments. Similarly, sleep-disordered breathing was not confirmed with polysomnography, the current standard for diagnosis of SDB. However, recent studies have found a strong correlation between a sleep score of >.33 and a diagnosis of SDB using polysomnography.30, 39 In addition, we did not have information regarding prior surgeries including tonsillectomy or adenoidectomy.
All families were recruited from an inner-city community and many of these families experience stressful lives that may contribute to parents’ report of both sleep and behavior problems. Fortunately, we are able to control for several factors including parent depression, stress and quality of life in our multiple regression analysis. In addition, this study uses data at the end of a 7–9 month asthma intervention, and while we were able to control for the influence of the intervention in our models, we realize that caregivers may respond differently to questions depending on their views of the intervention.
Lastly, it is possible that parents may confuse some symptoms of sleep-disordered breathing with nighttime asthma symptoms. For example, symptoms such as ‘struggling to breathe’ or ‘intermittent breathing’ at night could be interpreted as either asthma symptoms or SDB. However, symptoms of snoring are less likely to be confused with symptoms of asthma, and when we repeated our analysis using the individual subscales (snoring and sleepiness) we found similar, consistent relationships between SDB and behavior.
Childhood sleep disorders are often overlooked in the clinical setting, and this study identifies a group of children who may be at particularly high risk. In 2002, the American Academy of Pediatrics recommended that physicians screen all children for snoring to determine risk of OSA.56 This recommendation underscores the importance of sleep disorders and children’s health. The findings of this study suggest that clinicians should be particularly diligent about screening all children with asthma for SDB, and consider sleep disorders as a possible risk factor for behavior problems. Further investigation is needed to determine whether treatment of sleep disorders would help to decrease behavior problems in this population.
This study was funded by the National Heart Lung and Blood Institute (R01-HL079954) and the Halcyon Hill Foundation
We would like to thank Kelly Conn, MPH for her assistance with this manuscript.