Aims Obstructive sleep apnoea syndrome (OSAS) is a common respiratory disorder of childhood, affecting 1–3% of normal children, rising to 13% in the morbidly obese and has even higher prevalence in children with co-morbidities. Polysomnography (PSG), is considered the ‘gold standard’ for diagnosis. PSG is a highly technical physiological monitoring system performed in a sleep laboratory and not widely available in the UK. Clinicians still rely upon clinical history and examination to diagnose and subsequently manage OSAS. The aims of this study were to determine the utility of clinical signs and symptoms in diagnosing OSAS and to evaluate their use in predicting the severity of OSAS.

Methods Retrospective study of 423 paediatric patients who underwent polysomnography for suspected OSAS. Patients were stratified according to diagnosis based on PSG findings; no OSAS (n=126), mild (n = 120), moderate (n = 85) or severe (n = 92). We analysed their sleep history ‘questionnaire’ containing information about sleep behaviour, daytime symptoms and physical examination. Sensitivity, specificity, negative and positive predictive values, odds ratios and likelihood ratios were used to determine the diagnostic accuracy of single and combined signs and symptoms in identifying OSAS and predicting severity.

Results No single symptom had both, high sensitivity and specificity in predicting paediatric OSAS. However, parental reported snoring, nocturnal mouth breathing and blocked nose were sensitive but not specific. Additionally, a ‘blocked’ nasal airflow finding on examination was highly associated with a diagnosis of OSAS (100% specificity). On analysis of OSAS severity, severe OSAS could be confidently excluded in the absence of snoring (LR- <0.1). Moreover, the presence of grade IV tonsils was a good predictor of severe OSAS (93.06% specificity). Five symptoms combined; snoring, snoring every night, witnessed apnoeas, nocturnal mouth breathing and blocked nose gave the best predictive model. This was specific (88.24%) in identifying OSAS subjects and sensitive (95.83%) in excluding a diagnosis of severe OSAS.

Conclusion The use of single and combined signs and symptoms has an important role in aiding the diagnosis and predicting the severity of paediatric OSAS. But clinical assessment alone is insufficient to reliably diagnose OSAS in lieu of overnight polysomnography.