OBJECTIVE. Childhood sleep-disordered breathing has an adverse impact on cognitive development, behavior, quality of life, and use of health care resources. Early viral infections and other immune-mediated responses may contribute to development of the chronic inflammation of the upper airway and hypertrophic upper airway lymphadenoid tissues underlying childhood sleep-disordered breathing. Breastfeeding provides immunologic protection against such early exposures. Therefore, we sought to explore whether sleep-disordered breathing severity would differ for children who were breastfed as infants.
METHODS. The parents or guardians of 196 habitually snoring children (mean ± SD: 6.7 ± 2.9 years old) who were undergoing overnight polysomnography at Kosair Children's Hospital Sleep Medicine and Apnea Center completed a retrospective survey on the method(s) used to feed the child as an infant.
RESULTS. Among habitually snoring children, those who were fed breast milk for at least 2 months had significantly reduced sleep-disordered breathing severity on every measure assessed, including apnea-hypopnea index, oxyhemoglobin desaturation nadir, and respiratory arousal index. Breastfeeding for longer than 5 months did not contribute additional benefits.
CONCLUSIONS. Our findings support the notion that breastfeeding may provide long-term protection against the severity of childhood sleep-disordered breathing. Future research should explore mechanism(s) whereby infant-feeding methods may affect the pathophysiology of development of childhood sleep-disordered breathing.
Sleep-disordered breathing (SDB), a condition marked by intermittent hypoxia and hypercapnia and sleep fragmentation, is present in 1 to 3% of children.1, 2 Children with SDB are more frequent users of health care services,3 experience more frequent comorbid chronic illnesses,4, 5 and may develop significant clinical cardiovascular morbidity.1, 6–9 They also display significant behavioral and cognitive comorbidities,10 as well as reduced quality of life and mood alterations.11 Animal models of SDB display long-term, partially irreversible neurocognitive consequences after intermittent hypoxia during sleep, particularly when such exposures occur early during development.12–14
The relatively high rate at which SDB is diagnosed among otherwise healthy children and the implications for untreated SDB are a major health concern, to the extent that the American Academy of Pediatrics recommended that primary caregivers routinely screen their pediatric patients for the presence of snoring,15 the cardinal symptom of SDB.16 The preferential treatment of obstructive sleep apnea in otherwise healthy children is surgical removal of adenoids and tonsils.17–20
Although the pathophysiologic mechanisms underlying hypertrophy of these upper airway lymphadenoid tissues has remained unclear, direction can be taken from current studies of the lower airway suggesting that early exposure to respiratory syncytial virus infection and other frequent respiratory viral pathogens may lead to potential neurogenic inflammation21 and airway remodeling, both of which may contribute to development of childhood asthma.22–24 In the United States, nearly all children are infected with respiratory syncytial virus by their third year.25 Preliminary evidence for the presence of similar pathways in development of upper airway lymphadenoid tissue hypertrophy has recently emerged.26, 27 Furthermore, inflammatory markers have been found in the adenotonsillar tissue of children who undergo surgery for SDB, and treatment with anti-inflammatory medications significantly reduces the respiratory parameters on polysomnography in these children.28
Breast milk and colostrum provide the infant with immunoglobulin A29, 30 and are considered a primary method for prevention of respiratory viral infections.31 However, the impact of breastfeeding practices on the severity of SDB in children has not been heretofore examined. We therefore sought to test the hypothesis that children who had been fed breast milk as infants would be less likely to develop significant childhood SDB compared with children who had not been fed breast milk as infants and that the duration of breastfeeding may affect such association.
Parents of habitually snoring children who were undergoing overnight polysomnography as either a clinic patient or a participant in a larger research study at the Kosair Children's Hospital Sleep Medicine and Apnea Center were invited to fill out a brief survey about the feeding method(s) that they used when their child was an infant. A cover letter that described implied consent and Health Insurance Portability and Accountability Act authorization was administered as approved by the institutional review boards at the University of Louisville and Norton Healthcare.
The paper-and-pencil survey included child demographics (age, gender, ethnicity, and birth order), whether anyone in the household smoked, and the relationship of the survey respondent to the child. The respondent was asked to identify the method(s) used to feed the child during the first year of life: formula only (no breast milk), breast milk only (no formula), or both formula and breast milk. For those who were fed any breast milk, the age in months at which the child was completely weaned was also identified. Finally, the respondent was asked, “How well do you remember this information?” on a 5-point scale from “not at all confident” to “completely confident.”
Polysomnography and Scoring
With the parent or guardian present, overnight polysomnography was performed by using commercially available multichannel data-acquisition equipment and scoring software (MedCare Diagnostics, Amsterdam, Netherlands) to record 8 channels of electroencephalography (O1/O2, P3/P4, C3/C4, F3/F4), submental chin and anterior tibialis electromyography, bilateral electrooculogram, snore sensor, electrocardiogram, chest and abdominal inductance plethysmography, pulse oxygen saturation (Spo2) and wave form, and oronasal airflow (assessed through an oronasal thermistor, end-tidal CO2, and nasal pressure transducer). Simultaneous video and audio monitoring was digitally recorded. No study was performed on a night when a child had an acute illness, fever, or nasal discharge. Records were scored by an analyst who was blinded to the infant-feeding–method status.
Stage scoring was performed using standard criteria.32 Because criteria for arousals have not yet been established for children, arousals were defined as recommended by the American Sleep Disorders Association Task Force report33 and manually scored as spontaneous or respiratory related (occurring immediately subsequent to an apnea, hypopnea, oxyhemoglobin desaturation, or snore) and reported as indices on the basis of occurrence per hour of total sleep time (TST).
Central apneas were scored on the basis of cessation of oronasal flow and chest wall and abdominal movement; obstructive apneas were scored in the absence of oronasal airflow with continued chest wall and abdominal movement, whereas decreases in oronasal flow ≥50% with continued effort were scored as hypopneas, provided that these events had a minimum duration of 2 breath-lengths and an associated ≥4% Spo2 desaturation and/or arousal.34, 35 Snoring was scored in the presence of a change in basal snore sensor levels that was additionally verified and annotated by the technologist via both in-room checks and microphone transmission. Spo2 nadir and oxyhemoglobin desaturations were calculated from valid Spo2 tracings during sleep with values during movement artifact excluded. Scoring of all respiratory events and oxyhemoglobin desaturations were initially automated and then user verified. Respiratory events were scored as indices on the basis of the number of events per hour of TST. An apnea-hypopnea index (AHI) was calculated on the basis of the number of combined apneas and hypopneas per hour of TST.
Descriptive statistics were calculated. Group comparisons with dichotomous variables were made by using the χ2 test; those with continuous variables were performed by using 1-way analysis of variance. Data were analyzed by using SPSS 14.0 (SPSS, Chicago, IL), and P < .05 was considered statistically significant. Cohen's d with pooled SDs was calculated to determine effect sizes.
Of the 212 surveys completed, polysomnography data were available for 196 children. There were no differences on demographic measures between habitually snoring children who were recruited from the larger research study (n = 89) and clinical patients (n = 107); therefore, their databases were combined. Child demographics are shown on Table 1. Surveys were completed by the child's mother (95.9%), father (2.6%), or grandparent (1.0%). The confidence with which the respondent remembered the information was reported at 3.20 (±0.93 [SD]) on a 5-point scale, where 0 = not at all confident and 4 = completely confident. Five respondents (2.6% of sample) reported a 0 or 1 on the recall question; removal of data from these low-recall respondents did not alter any analysis, and they were retained.
When they were infants, 52% of the children were fed formula only, 10% were fed breast milk only, and 38% were fed a combination of formula and breast milk. For those who were fed breast milk in any amount, the average age of weaning was 7.3 ± 7.0 months. No age, gender, ethnicity, or household environmental tobacco smoke exposure differences were found on feeding method, duration of breastfeeding, or polysomnography measures. Group polysomnography measures are shown in Table 2.
Two group comparisons were made. First, comparisons were made between infants who were never breastfed and those who were breastfed for increasing durations up to ≥12 months (eg, 0 vs ≥1 month of breastfeeding, 0 vs ≥2 months of breastfeeding). Compared with children who were never fed breast milk, those who were fed breast milk had significantly lower AHI at every duration of breastfeeding. The breastfed children also had higher Spo2 nadir through 11 months of breastfeeding and lower respiratory arousal indices than those who were never breastfed.
Second, comparisons were made between infants who were breastfed for increasing durations compared with those who were breastfed for less than that duration (eg, <2 vs ≥2 months of breastfeeding, <3 vs ≥3 months of breastfeeding) There was an insufficient number of children who breastfed <2 months for comparisons. (Note that only breastfed children were included in these comparisons, and annotation of breastfeeding duration “<2 months” does not include 0 months.) Longer duration of breastfeeding was associated with lower AHI through 4 months of breastfeeding, higher Spo2 nadir through 5 months of breastfeeding, and lower respiratory arousal index through 3 months of breastfeeding. Breastfeeding beyond 5 months did not seem to provide added benefits on these measures of SDB compared with infants who had been breastfed for shorter durations.
Group means, variance, and statistical differences for comparisons between no breastfeeding and increasing durations of breastfeeding as well as between breastfeeding of different durations are shown in Fig 1 for AHI, Fig 2 for Spo2 nadir, and Fig 3 for respiratory arousal index. There were no spontaneous arousal index differences between groups on the basis of any duration of breastfeeding. The number of children who were breastfed for successive durations and the effect sizes for statistically significant differences on each polysomnography measure comparison are shown in Table 3.
This study shows that among children who had symptoms of SDB, those who were fed breast milk when they were infants had significantly reduced disease severity. Breastfeeding for >5 months did not seem to contribute additional benefits. The differences among polysomnography measures of SDB between children who had and had not been breastfed were clinically significant: children who were breastfed had polysomnography-related severity measures below the standard threshold for surgical treatment.
We see 2 possible explanations for these findings. First, as described in the introduction, the reduced severity of SDB in breastfed children may be explained by the immunologic protection that is provided by breast milk. This notion is supported by the growing evidence suggesting that early respiratory viral exposures may contribute to the development of childhood SDB by facilitating airway remodeling and, potentially, promoting propitious conditions for enhanced proliferation of upper airway lymphadenoid tissues ultimately facilitating the occurrence of adenotonsillar hypertrophy.21, 23, 24, 27, 28 Indeed, overgrowth of these upper airway tissues is 1 of the major causative factors of pediatric SDB, such that surgical removal of these tissues is the first-line treatment for SDB in otherwise healthy children without craniofacial or neuromuscular abnormalities.17–20, 36 It is likely that in the context of differential exposures and immune responses to respiratory pathogens during early infancy as modulated by breastfeeding practices, differences would also emerge in the immunologic substrate of upper airway lymphadenoid tissues. In support of such assumption, adenotonsillar tissues of children with SDB are distinct from those with recurrent infections in regard to the expression and distribution of cysteinyl leukotriene receptors, suggesting that the mechanism for inflammation may be different in children with SDB.37
Another possible explanation for our findings is that oral cavity features such as high palates, narrow dental arches, and retruded chin all are additional risk factors for SDB in children.38 Breastfeeding promotes healthy jaw formation, thereby preventing the occurrence of many of these anatomic issues39, 40; therefore, the mechanical aspects of breastfeeding may provide additional protection against development of SDB in that this feeding method promotes an upper airway that is less vulnerable to narrowing and collapse during sleep.
It is important to note that although our findings indicate that disease severity is reduced in association with breastfeeding, our work and that of others have consistently shown that even clinically mild levels of childhood SDB are associated with cognitive and behavioral compromise (see review by Beebe10). Our work should not be interpreted to suggest that breastfeeding entirely prevents the development of SDB. Although severity was shown to be reduced in this sample, there is no consensus on what constitutes a safe level of SDB, if any, so treatment should not be delayed on the basis of a child's feeding method.
There are several limitations to this preliminary investigation of the relationship between infant feeding and development of SDB. First, the retrospective feeding-methods survey has not been validated, although the high rate of confidence with which the respondent recalled the information was reassuring. Second, the proportion of breast milk feedings that were given by bottle is unknown and, as described, this may have some bearing on the mechanism by which breastfeeding has its beneficial effects on severity of SDB. Future investigations of this effect may use a dosage-response approach to take into account the proportion of breast milk feedings given via breast versus bottle and the type of nipple used for bottle feedings. Furthermore, although we did not find ethnicity-related differences on the incidence or duration of breastfeeding in this sample, future work should acquire more precise data to control for socioeconomic status as a possible confounder. Data on familial history of snoring and presence of SDB should also be collected in future investigations. Finally, because all of the children in this sample were habitual snorers, we did not measure differences in snoring prevalence between former breastfed and formula-fed infants. Future investigations into differences in rates and severity of SDB on the basis of infant-feeding methods is necessary.
Our findings support the notion that breastfeeding may provide long-term protection against the incidence and/or severity of childhood SDB. Future research aiming to explore the mechanism(s) by which infant-feeding methods may affect the pathophysiology of childhood SDB seems warranted.
This study was funded by National Institutes of Health grants F32 HL-074591 (to Dr Montgomery-Downs) and HL65270 (to Dr Gozal).
We are grateful to the study participants. Jennifer Bruner, RPSGT, Nina Burkhead, Karen Conrad, Janie Cook, David Davis, Renee Ferguson, Darlene Herps, RPSGT, Contessa Howlett-Tyler, Joseph Lipcovich, Tamela Nichols, Dianna O'Neal, Allison Parker, Julie Render, Andrea Shewmaker, Nigel Smith, RPSGT, Tonya Thornton-Gallahar, Nicky Wilkerson, Lisa Witcher, and Sarah Wheeler-Bayens assisted with sleep-data collection.
- Accepted May 17, 2007.
- Address correspondence to Hawley Evelyn Montgomery-Downs, PhD, West Virginia University, Department of Psychology, 1124 Life Sciences Building, PO Box 6040, Morgantown, WV 26506-6040. E-mail:
The authors have indicated they have no financial relationships relevant to this article to disclose.
Dr Crabtree's current affiliation is Division of Behavioral Medicine, St Jude Children's Research Hospital, Memphis, TN.
- ↵Reuveni H, Simon T, Tal A, Elhayany A, Tarasiuk A. Health care services utilization in children with obstructive sleep apnea syndrome. Pediatrics.2002;110 :68– 72
- ↵Rona JR, Li L, Gulliford MC, Chinn S. Disturbed sleep: effects of sociocultural factors and illness. Arch Dis Child.1998;78 :20– 25
- ↵Schechter MS; American Academy of Pediatrics, Section on Pediatric Pulmonology, Subcommittee on Obstructive Sleep Apnea Syndrome. Technical report: diagnosis and management of childhood obstructive sleep apnea syndrome. Pediatrics.2002;109(4) . Available at: www.pediatrics.org/cgi/content/full/109/4/e69
- ↵American Academy of Pediatrics, Section on Pediatric Pulmonology, Subcommittee on Obstructive Sleep Apnea Syndrome. Clinical practice guideline: diagnosis and management of childhood obstructive sleep apnea syndrome. Pediatrics.2002;109 :704– 712
- ↵Kimpen JLL, Simoes EAF. Respiratory syncytial virus and reactive airway disease; new developments prompt a new review. Am J Respir Crit Care Med.2001;163(3 pt 2) :S1
- ↵Ogra PL. Respiratory syncytial virus: the virus, the disease and the immune response. Paediatr Respir Rev.2004;5(suppl A) :S119– S126
- ↵Goldbart AD, Mager E, Veling MC, et al. Neurotrophins and tonsillar hypertrophy in children with obstructive sleep apnea. Pediatr Res.2007; In press
- ↵Tsutsumi H, Honjo T, Nagai K, Chiba Y, Chiba S, Tsugawa S. Immunoglobulin A antibody response to respiratory syncytial virus structural proteins in colostrum and milk. J Clin Microbiol.1989;27 :1949– 1951
- ↵Hanson LA, Korotkova M, Telemo E. Breast-feeding, infant formulas, and the immune system. Ann Allergy Asthma Immunol.2003;90(suppl 3) :59– 63
- ↵Rechtschaffen A, Kales A, eds. A Manual of Standardized Terminology, Techniques and Scoring System for Sleep Stages of Human Subjects. Los Angeles, CA: Brain Information Services/Brain Research Institute, University of California; 1968
- ↵Montgomery-Downs HE, O'Brien LM, Gulliver TM, Gozal D. Polysomnographic characteristics in normal preschool and early school-age children. Pediatrics.2006;117 :741– 753
- Copyright © 2007 by the American Academy of Pediatrics