Correlations between polysomnographic and lateral airway radiograph measurements in paediatric obstructive sleep apnoea

Aim To evaluate the ability of lateral airway radiography (LAR) to assess adenoidal hypertrophy in children and correlate with the severity of obstructive sleep apnoea (OSA). Methods This cohort study was undertaken in 72 children who presented consecutively for evaluation of OSA to the outpatients of the Children's Hospital at Westmead. All children had LAR and overnight polysomnography (PSG). Five assessors, with varying experience, were blinded to the PSG results and independently analysed the LAR. Inter-rater reliability was determined for four published assessment methods; Hibbert, Johannesson, Fujioka and Cohen and Konak. We then compared the four LAR results with PSG-determined criteria for OSA. Results Using intraclass correlations, inter-rater correlations were moderate to high for all four standardised evaluations of LAR with values ranging from 0.51 to 0.96. With the radiologist taken as the 'gold standard', individual assessors ranged from 0.05 to 0.91. LAR correlated best with PSG determined obstructive apnoea hypopnea index and minimum oxygen saturation for the anterior airway measurement (Hibbert) with r-values of -0.25 and 0.25 respectively (P < 0.05). Conclusion Amongst four methods of evaluating adenoid size, the anterior airway size correlated best with PSG variables of obstructive respiratory index and minimum oxygen saturation. However, the methods are not able to be used as a predictor for OSA.