Objective. Our objective was to investigate the occurrence of snoring in a general population sample of children and to evaluate the association with anthropometric data and clinical findings of oropharynx and nasal airways.
Methods. A cross-sectional study was conducted with children from primary and secondary schools in Civitavecchia and Viterbo in the Latium region in central Italy. The total sample of the survey included 2439 schoolchildren. A total of 2209 children who were ages 10 to 15 years were selected (response rate: 90.5%) according to their snoring frequency during sleep: never, only with colds, occasionally apart from with colds, often. Children in the last category were defined as habitual snorers. Data were collected by means of questionnaires and clinical examination. A blood sample was collected to determine the concentration of hemoglobin in the blood. Logistic regression was used to calculate odds ratios (OR) and 95% confidence intervals.
Results. The prevalence of habitual snorers was 5.6%. Boys who were older than 15 years and had a body mass index greater than the 90th percentile were significantly more likely to be snorers. Habitual snoring was strongly associated with decreased nasal patency (rhinitis OR: 2.13; septal deviation OR: 2.75; nasal obstruction OR: 2.20). Children who had undergone adenoidectomy or had markedly enlarged tonsils were at greater risk of being habitual snorers (OR: 4.28 and 5.07, respectively). Last, habitual snorers had a significantly higher concentration of hemoglobin in the blood compared with other children.
Conclusion. Body weight and nasal and pharynx patency seem to be the main determinants of snoring. The finding of higher values of blood hemoglobin concentration in snorers than in nonsnorers suggests that these children could be experiencing oxyhemoglobin desaturation during sleep. Taking into consideration the relationship between these different risk factors could lead to a better clinical approach to the snoring child.
Although snoring often is regarded as a secondary phenomenon, it may indicate the presence of sleep apnea or a lesser form of sleep-disordered breathing. The prevalence of snoring could vary between 3.2% and 12%1–4 depending on age, and a recent review estimated that snoring with sleep-related daytime symptoms may occur in approximately 11% of children.5Daytime sleep symptoms include daytime sleepiness as well as various behavioral problems, such as inattention, hyperactivity, and aggression.3 In a 2-year study, Ali et al6found that habitual snoring and behavioral problems tended to disappear spontaneously in approximately half of the children investigated, whereas sleepiness, hyperactivity, and restless sleep were still present in persistent snorers and among incident cases. Snoring children could have significantly higher oxygen saturation dip rates than control subjects3 as snoring may be the first step toward the sleep apnea syndrome with hypoxic respiratory events during sleep. However, the clinical history may fail to distinguish primary snoring from obstructive sleep apnea syndrome in children,7 and parents might be unaware of the problem.8 Risk factors for snoring are considered to be obesity,5 upper airway obstruction, which is commonly associated with nasal obstruction,9 or hyperplasia of adenoid and/or tonsillar tissue.10
We investigated the occurrence of snoring in a general population sample of children and evaluated the association with anthropometric data and with clinical findings of oropharynx and nasal airways.
A survey of respiratory disease in a population sample of schoolchildren who lived in 2 areas in the Latium region, Italy (Civitavecchia, urban area, and 3 rural communities of the province of Viterbo) was conducted in 1990 to 1991. Details of the study design have been reported previously.11 All children who attended the fifth grade of 4 primary schools, which were randomly chosen in various communities of the Latium region, and all adolescents who attended the nearby 6 secondary schools were invited to participate. The overall selected sample comprised 2593 children, whereas the total sample examined in the 2 years of the survey included 2439 children. Parents were asked to complete a self-administered questionnaire adapted from the American Thoracic Society questionnaire on children.12 Snoring was investigated with the following question: “Does your child ever snore?” On the basis of the answer, 4 categories were formed: 1, never; 2, only with colds; 3, occasionally without a cold; and 4, often. Children in the last category were defined as habitual snorers. Asthma was defined as either physician-diagnosed asthma (American Thoracic Society question 29A) or as at least 3 of the following: wheezing with colds (question 17A), wheezing without a cold (question 17B), dyspnea associated with wheezing (question 18A), and wheezing after exercise (question 19).
Our questionnaire included questions on rhinitis (“Has your child ever had rhinitis [stuffy and running nose, itchy eyes, sneezing]?”), otitis (“Has your child ever had otitis”), and exposure to parental smoking. Parents also were asked whether their child had undergone a tonsillectomy and/or adenoidectomy. Socioeconomic status was evaluated by considering father's education (<6 years, 6–8 years, 9–13 years, and 13+ years, which correspond to years spent at primary school, secondary school, high school, and university, respectively). The number of inhabitants of the dwelling and house size (number of rooms) were determined, and household crowding was calculated by dividing the number of inhabitants by the number of the rooms (low, <1; medium, 1–2; high, >2). Last, the child was interviewed confidentially by a physician about personal active cigarette smoking (never, sometimes <1 cigarette/wk, at least 1 cigarette/wk). An informed consent was required from parents before clinical examinations and blood test were performed.
Medical examination was performed by a physician together with measurement of weight (kg) and standing height (cm) in stocking feet. The body mass index (BMI; weight/height2) was calculated and divided using standardized percentile curves13 in 4 classes: up to 75th percentile, 76th to 90th percentile, 91st to 95th percentile, and 96th to 100th percentile. The nose was examined to evaluate the presence of nasal obstruction or the presence of septal deviation. The size of the tonsils was estimated by examination of the throat and scored on a 4-point scale: 1, normal-size (not visible or extending to the pillars); 2, absent because of surgery; 3, moderate enlargement (enlarged beyond the pillars but not meeting the uvula); 4, marked enlargement (meeting the uvula or “kissing” at the midline).
During the last 5 months of the survey, the children were asked to give blood samples for a number of investigations, including the measurement of blood hemoglobin. Venous blood samples were drawn from 997 of 1220 children (82%).
The data analysis was conducted in various stages. First, prevalence of different categories of snoring for each variable were analyzed and results were given as percentages, and the linear trend was analyzed by χ2 test for trend. Various factors were evaluated: gender, age, BMI, father's education, household density, parental smoking, asthma, rhinitis, otitis, septal deviation of the nose, nasal obstruction, adenoidectomy, and dimension of tonsils were the risk factors. Second, the role played by different risk factors in determining snoring was examined by polytomous logistic regression analysis. Snoring habits (only with colds, without a cold, and habitual snorers) were analyzed using nonsnorers as reference category. Results were given as odds ratios (OR) with 95% confidence intervals (95% CI) Last, the relationship between snoring and concentration of hemoglobin in the blood was analyzed by a linear regression analysis using concentration of hemoglobin as the dependent variable and age, gender, BMI, height, father's education, parental smoking, and habitual snorers (as dummy variable) as independent variables. All of the analyses were performed using Stata Statistical Software (release 6.0; Stata Corp, College Station, TX).
Questionnaires and clinical examination were completed for 2209 children, representing an overall response rate of 90.5%.
Tables 1 and 2 show the association between snoring and the demographic and personal characteristics of the sample studied. On the whole, 124 children (5.6%) were found to be habitual snorers and 229 (10.4%) were reported to snore without having a cold. Boys were significantly more likely than girls to be habitual snorers (OR: 1.55). Although snoring did not change linearly over the age categories, children who were age 15 years or older had a significantly higher risk of being habitual snorers (OR: 3.32). Snoring increased significantly in children with BMI >90th percentile (χ2 for trend:P = .000). Moreover, the prevalence of snoring was more than double in children with a BMI >95th percentile compared with children with BMI >75th percentile (OR: 2.66). No statistically significant association was found between snoring and parental education or household crowding, although habitual snorers tended to be more prevalent in more crowded families (χ2 for trend: P = .007; Table 2). Children whose parents smoked were more likely to be snorers than children whose parents did not smoke (OR: 1.58; χ2 for trend:P = .004). Last, snoring increased in active smokers (10.2% vs 5.2%; P = .058 in univariate analysis), but this association was far from statistical significance when adjusted for other covariates, as smoking habits of adolescents were strongly associated with smoking habits of parents (χ2for trend: P = .0025).
Table 3 shows the association of snoring with clinical factors. Children who had asthma were more likely to snore with colds, without a cold, and habitually, even if the OR within the single categories failed to reach statistical significance. Rhinitis was strongly associated with snoring in all of the categories in which children with rhinitis had a higher risk of snoring with colds (OR: 2.24), without a cold (OR: 1.93), and habitually (OR: 2.13). Prevalence of snoring was increased in children who had had otitis, with a significant association between otitis and habitual snoring (OR: 1.65). A previous adenoidectomy was strongly associated with snoring only with colds (OR: 1.64), without a cold (OR: 1.92), and habitually (OR: 4.28). The clinical examination showed that habitual snoring was more frequent in children with a septal deviation of the nose, nasal obstruction, and a marked enlargement of the tonsils. The ORs were 2.75, 2.20, and 5.07, respectively. Tonsillectomy increased the risk of snoring only with colds (OR: 1.55), whereas no association was found regarding the presence of snoring without colds or habitually.
The mean value of hemoglobin was 13.8 g/dL (±1.11), and adjusting for age, gender, BMI, height, father's education, and parental smoking, habitual snorers had a value of hemoglobin greater than other children's (14.14 g/dL [standard error: 0.14] vs 13.74 g/dL [standard error: 0.04]; P = .006).
Our study demonstrated that the main predictors of snoring in adolescents are increased body weight, a decreased nasal patency (rhinitis, septal deviation, nasal obstruction), and clinical conditions that could narrow the pharyngeal airway (tonsils hypertrophy, adenoidectomy without tonsillectomy). The increased hemoglobin concentration in snorers suggests that some of them could experience hypoxia during sleep.
The prevalence of habitual snoring in our sample was 5.6%. Previous studies found a higher prevalence of snoring (7.6%2 to 10%1) in the younger age group, the prevalence being the same in boys and in girls. Moreover, the prevalence tended to decrease with age2 because of an increase in the pharyngeal cross-sectional area with growth. However, a recent survey in young adults reported that snoring occurred in approximately 25% of the sample with a higher prevalence in boys than in girls.14We found that habitual snoring occurred more frequently in boys than in girls and that habitual snorers were more likely to be older with a significant OR in children age 15 years. Although the formal test of interaction between gender and age was not significant, when we repeated the analysis separately for boys and girls, the risk of being habitual snorers was reduced and not significant in girls age 15 years (OR: 1.91; CI: 0.4–8.8), whereas it had increased in boys (OR: 4.23; CI: 1.6–11.2). This suggests that the increased levels of testosterone in boys after puberty, which has an apnea-promoting effect, could increase the prevalence of snoring in boys.
Obesity is a widely known risk factor for sleep-disordered breathing in children as was demonstrated in a community-based epidemiologic study of sleep apnea.15 We found that a BMI >90th percentile was associated with snoring, and the association was more relevant in children with a BMI >95th percentile. Rosen16demonstrated that although obesity was not a predictor of obstructive sleep apnea syndrome, it was associated with snoring and transient oxyhemoglobin desaturation events during sleep. Furthermore, respiratory problems during sleep have been found to increase in obese children when adenoid and/or tonsil enlargement was present.17
Social class (ie, level of paternal education) did not seem to influence significantly the prevalence of snoring in the overall sample, although habitual snoring seemed to be less common in children whose father had a high level of education. Household crowding seemed to influence the prevalence of snoring and could be considered an indicator of social class, and it also has been demonstrated as a risk factor in increased exposure to passive smoking,11 which can, in turn, lead to an increase in the prevalence of snoring as has been demonstrated in previous studies.2,,3
Our data confirm passive parental smoking as risk factor for snoring in children.2 As snoring depends on pharyngeal size, smoking can provoke mucosal edema and inflammation, resulting in a narrowing of the pharynx. As smoking habits of children were strongly associated with smoking habits of parents, we were unable to find an association between active smoking and snoring in our sample, although this association was present in the univariate analysis. It has been demonstrated that active smoking is a widely known risk factor for habitual snoring18 and sleep-disordered breathing19 in adults. This finding has been confirmed in young students.20 This strong association has lead to the hypothesis that sleep apnea could be a predisposing factor for tobacco use in adolescents.21 Because the main symptoms of sleep apnea are somnolence and obesity, nicotine could decrease these 2 symptoms as well as reduce the frequency and duration of apneas; thus, apneic adolescents could use tobacco to reduce their symptoms. Conversely, some data suggest a causal role of cigarette smoking in promoting sleep disturbance.18,,22The relation between tobacco and sleep-disordered breathing deserves additional investigation.
Recent data suggested that a history of asthma is associated with sleep-disordered breathing.15 In our data, although snoring was significantly increased in children with asthma, the statistical significance decreased when asthma was included in the logistic regression, suggesting that factors such as rhinitis and nasal obstruction, which were more common in children with asthma than in other children (20.8% vs 6.7% and 12.1% vs 8%, respectively), could be the determinants of snoring in children with asthma.
In our sample, both nasal deviation and nasal obstruction as well as the presence of rhinitis were found to be strong predictors of snoring. Nasal influences on snoring and sleep apnea are widely known.9 The knowledge of some features of nasal physiology could help in understanding the role of a partial nasal obstruction in the pathogenesis of snoring. Nasal resistances represent approximately half of the total respiratory resistances in humans, and changes of vascular tone (vasoconstriction or vasodilatation) could modify the nasal mucosa so as to increase or to reduce the nasal patency. The nasal patency is a dynamic phenomenon because an alternating congestion and decongestion of the nasal airway (ie, “nasal cycle”) occurs in approximately 70% of the adult population. In addition, nasal congestion can vary with changes in body and head position with an increase in nasal resistance when the individual is supine. However, the lateral recumbency decreases the patency of the ipsilateral nasal passage and increases the patency of the contralateral nasal passage. Last, nasal congestion is likely to be present in individuals with rhinitis. An individual with a fixed asymptomatic nasal obstruction, either unilateral or bilateral, may easily experience an increase in total airway resistance because of the lateral recumbency or as a result of the nasal cycle.9,,23 It has been demonstrated that nasal obstruction leads to disturbed sleep, apneic episodes, and loud snoring.24 A recent population-based study in adults showed that individuals with nasal congestion attributable to allergy or nasal obstruction were more likely to be habitual snorers with frequent episodes of apnea and hypopnea during sleep.25Otitis was found to be associated with snoring, and it has been suggested that recurrent ear infections could lead to suspect sleep-disordered breathing, particularly when associated with throat infections.5
Adenoidectomy without tonsillectomy was found to be a risk factor for snoring. This finding confirms that adenoidectomy alone could be insufficient treatment. Moreover, although adenoidectomy is one of the surgical procedures of choice when symptoms of sleep-disordered breathing are present, symptoms could still be present in some patients26 after surgery, possibly because adenoidal hypertrophy during growth could lead to an anatomic change of the pharynx. A recent study of snoring among children demonstrated that adenoidectomy was a risk factor for obstructive sleep apnea syndrome.27
As expected, tonsil hypertrophy was strongly related to snoring. Moreover, children with enlarged tonsils showed a greater BMI than children with normal tonsils (P = 0.001 for trend, adjusted for age and gender); normal children had a BMI of 20.50 versus a BMI of 20.96 in children with moderately enlarged tonsils and a BMI of 21.42 in children with markedly enlarged tonsils. To our knowledge, there are no data on the association between obesity and tonsillar hypertrophy, although a study in obese children with sleep-disordered breathing reported that tonsillar hypertrophy occurred in 59.3% of children.17
The finding of higher values of blood hemoglobin concentration in snorers than in nonsnorers suggests that these children could be experiencing oxyhemoglobin desaturation during sleep. Oxyhemoglobin desaturation in children with sleep-associated breathing disorders has been described in select samples10,,17 and population surveys.3,,4 Obesity and adenotonsillar hypertrophy have been suggested to be predictors of the lowest oxyhemoglobin saturation during sleep in children with obstructive sleep apnea.10Last, although operations (ie, adenoidectomy and tonsillectomy) were found to be remarkably effective in treating sleep-associated breathing disorders in overweight children with adenotonsillar hypertrophy, it has been demonstrated that the reduction of body weight could improve further the oxyhemoglobin saturation during sleep.28
Our data add additional evidence to clinical investigation of snoring in children. In addition, because snoring could be located in the gray zone between physiology and pathology and the clinical history could fail to distinguish primary snoring from obstructive sleep apnea syndrome in children,7 recognizing and then treating 1 or more risk factors could both improve the quality of sleep and prevent the onset of a sleep apnea syndrome. In Fig 1, we summarize the distribution of the main risk factors such as obesity (BMI >90th percentile), nose problems (septal deviation, nose obstruction, rhinitis), and pharynx problems (adenoidectomy, hypertrophy of tonsils, otitis). Whereas the prevalence of a single risk factor did not differ between snorers and nonsnorers, presence of 2 or 3 risk factors was more than double in snorers compared with nonsnorers (ie, obesity + pharynx problems = 11.3% in habitual snorers vs 4% in nonsnorers). Because evaluation of body weight, nose, and pharynx is common in a clinical setting, we believe that taking into consideration the relationship among these different risk factors could lead to a better clinical approach to the snoring child.
- Received March 12, 2001.
- Accepted June 14, 2001.
Reprints requests to (G.M.C.) Respiratory Physiology Department, Catholic University, Largo F. Vito 1 00168 Rome, Italy. E-mail:
- BMI =
- body mass index •
- OR =
- odds ratio •
- CI =
- confidence interval
- Corbo GM,
- Fuciarelli F,
- Foresi A,
- De Benedetto F
- Ali NJ,
- Pirson DJ,
- Stradling JR
- Ali NJ,
- Pitson DJ,
- Stradling JR
- Olsen K,
- Kern EB
- Dell'Orco V,
- Forastiere F,
- Agabiti N,
- et al.
- Ferris BG
- Nieminen P,
- Tolonen U,
- Lopponen H,
- Lopponen T,
- Luotonen J,
- Jokinen K
- Kudoh F,
- Sanai A
- Copyright © 2001 American Academy of Pediatrics