Obstructive sleep apnea (OSA) is a clinical syndrome characterized by snoring, apneas and difficulty in breathing. These symptoms can be rated and a risk score (Brouillette score) can be calculated to estimate the likelihood of OSA. This study aimed at establishing the predictive value of the Brouillette score and observation by parents at home in children with syndromic or complex craniosynostosis, compared with ambulatory polysomnography. This prospective study included 78 patients (37 boys, mean age 7.3 years). Sensitivity and negative predictive values were calculated. Polysomnography showed clinically significant OSA in 11 children. The Brouillette score had a negative predictive value of 90% and a sensitivity of 55% in comparison with polysomnography. More than three-quarters of all patients snored. The single question ‘Is there difficulty with breathing during sleep?’ showed a sensitivity of 64% and a high negative predictive value of 91%. Thus, asking parents whether the child has difficulty in breathing during sleep can exclude the presence of clinical significant OSA and avoid polysomnography in children with syndromic and complex craniosynostosis.
Obstructive sleep apnea (OSA) is a clinical syndrome characterized by difficulty in breathing, snoring and apneas during sleep resulting in sleep fragmentation, hypoxia and hypercapnia. Other features of OSA are restless sleep, mouth breathing and sweating. The ‘gold standard’ for diagnosing the presence and severity of OSA is polysomnography (PSG), but a feasible alternative is a questionnaire about the presence of symptoms. After discriminant analysis, B rouillette et al. developed an OSA score, known as the Brouillette score, to predict the presence of OSA with high sensitivity . This score is calculated from a respondent’s rating on three items ( Fig. 1 ). Some studies showed that the Brouillette score could not reliably distinguish between the presence of OSA and simple snoring and that its sensitivity and specificity were not sufficient for affirming OSA .
Children with syndromic or complex craniosynostosis have a 40% risk of developing OSA due to midface hypoplasia and collapse of the pharynx. They must be screened for OSA from birth on. This is usually done by PSG, as the value of the Brouillette score as a screening tool in these children has not been established. An earlier study by the authors found a discrepancy between the high prevalence of OSA as established by the questionnaire and an analysis of the medical records . The present study aimed to determine the reliability of the Brouillette score and parents’ observation at home compared with ambulatory PSG to predict clinically significant OSA in children with syndromic or complex craniosynostosis.
Material and methods
A prospective cohort study was carried out at the authors’ hospital. All patients between 0 and 18 years with syndromic or complex craniosynostosis registered at the Dutch Craniofacial Center were invited to participate in the study between January 2007 and March 2008. Syndromic craniosynostosis included Apert, Crouzon, Muenke, Pfeiffer and Saethre-Chotzen syndromes. Complex craniosynostosis was defined as fusion of two cranial sutures or more without known fibroblast growth factor receptor (FGFR) or TWIST gene mutation 98 of the eligible 111 patients (88%) were included after informed consent.
The study at home had three components. The parents rated the three items of the Brouillette score (breathing difficulty, apnea and snoring) with regard to the sleep breathing pattern of their child over the previous 3 months. The parents observed their child at home during sleep for one period of 30 min and rated the items of the Brouillette score every 5 min and recorded any mouth breathing. The children underwent a cardiorespiratory PSG at home for one night.
The data were incomplete for 20 patients. Questionnaires on two patients and observation forms on 15 were not completed for various logistic reasons. During PSG, the total sleep time was below 360 min for three patients, which was too short for analysis. The data from 78 patients was analysed: 37 boys and 41 girls with a mean age of 7.3 ± 5.4 years (SD) at the time of PSG.
From the questionnaire and the observation form a Brouillette score (Br score 1) and an observation score (Br score 2) were calculated using the equation 1.42 D + 1.41 A + 0.71 S − 3.83 ( Fig. 1 ) . OSA is likely if the score is above −1 and is thought to be absent if the score is below −1. Mouth breathing was considered as continuous if parents observed mouth breathing during the whole observation.
Ambulatory PSG was carried out with the Embletta Portable Diagnostic System and analysed with Somnologica for Embletta software 3.3 ENU (Medcare Flaga, Reykjavik, Iceland). Thoracic and abdominal movements, nasal flow, saturation and pulse were monitored. A minimum of 360 min total sleep time was required. Obstructive apnea was defined as absence of airflow (measured by a nasal cannula) or as out-of-phase movement of thorax and abdomen (scored as X-flow). Hypopnea was defined as ≥50% reduction in nasal flow signal amplitude or X-flow signal amplitude, both for more than two breaths . The X-flow signal was the sum of the amplitudes of the thoracic and abdominal movements and was used when nasal airflow was insufficient. Mixed apnea was defined as a type of obstructive apnea with a central component that mostly preceded the obstructive pattern, for more than two breaths. Central apneas were not included in this study. Desaturation was defined as ≥4% decrease with respect to the baseline value. The severity of OSA was expressed in an obstructive apnea hypopnea index (OAHI), which consisted of: the hourly number of obstructive and mixed apneas; and the hourly number of hypopneas followed by desaturation. A score of ≤5 is considered to be of no clinical significance with no necessity to treat, between 6 and 25 as moderate OSA, and >25 as severe OSA .
For statistical analysis, contingency tables were made and the sensitivity (sens) and negative predictive value (NPV) with accessory 95% confidence intervals (CI) were calculated. The sensitivity of the questionnaire and observation (the number of Br scores ≥−1 that correctly identified OSA) was tested in comparison with the results of the PSG. The negative predictive value (the number of Br scores <−1 that correctly diagnosed the absence of OSA) of the two scores was calculated.
For 52 of the 78 patients (67%) the Brouillette score (Br score 1) was <−1. For 57 of the 78 patients (73%) the observation score (Br score 2) was <−1. Continuous mouth breathing was observed in 23 patients. The X-flow was used in 15 of them, for whom nasal flow registration was insufficient. Eleven PSGs were clinically significant and scored as OSA, based on OAHI.
The predictive results for the presence of clinical significant OSA are shown in Table 1 . The questionnaire had a high negative predictive value of 90% and a sensitivity of 55% when related to PSG. Combining the questionnaire with the parents’ observation gives a slight improvement of predicting OSA (sensitivity 64% and negative predictive value 91%). In 89% of the observations the findings of the parents for the observed 30 min corresponded with those for the matching PSG period.