Abstract

Aims: (1) To determine the extent to which tonsil size contributes to the severity of obstructive sleep apnoea (OSA) in children; and (2) to assess the use of tonsillar-pharyngeal (TP) ratio in differentiating patients with different severity of OSA.

Methods: Lateral neck radiograph was performed on 35 children referred consecutively to a university paediatric chest clinic for suspected OSA secondary to tonsillar hypertrophy. The tonsil size was determined by measuring the TP ratio on the radiographs. The severity of OSA was assessed by overnight polysomnography.

Results: A total of 24 boys and 11 girls (median age 6.2 years) were studied. All presented with symptoms of OSA, and tonsillar hypertrophy was detected on clinical examination. The median apnoea–hypopnoea index (AHI) was 16.93 (interquartile range: 8.41 to 28.29). The median TP ratio was 0.76 (interquartile range: 0.65 to 0.80). AHI was positively correlated with the TP ratio. The clinical tonsil size did not correlate with the AHI or the TP ratio. Using a TP ratio of 0.479, the sensitivity and specificity in predicting cases with moderate/severe OSA (AHI >10) were 95.8% and 81.8% respectively, while the positive and negative predictive values were 92.0% and 90.0% respectively.

Conclusions: Results show that in a population of children with OSA, tonsillar hypertrophy as assessed by lateral neck radiograph correlates positively with the severity of obstructive sleep apnoea. The TP ratio has high sensitivity and specificity in predicting those with moderate/severe disease and this feature may be used as a clinical screening method in prioritising patients with OSA for further assessment.