Can sleep questionnaires predict adenotonsillectomy outcome for children with sleep disordered breathing?

Introduction: Adenotonsillar hypertrophy is the main cause of childhood sleep disordered breathing (SDB) and adenotonsillectomy (TA) the most common treatment. Polysomnography (PSG) for diagnosing SDB is often difficult to obtain with Otolaryngologists usually relying on history and examination when recommending TA. Questionnaires assessing quality of life (QoL) may assist the Otolaryngologists decision making. Aims: To explore changes in QoL tools following TA for SDB in children aged 3 to 15 with the aim of identifying whether the Pediatric Sleep Questionnaire (PSQ) or Obstructive Sleep Apnoea -18 (OSA-18) is a better predictor of outcome following TA. Methods: QoL was assessed using OSA-18, PSQ and the Pediatric Quality of Life Inventory (TM) (PedsQL (TM)). Four groups were recruited from three research databases, those with: SDB, recurrent tonsillitis (RT), SDB and RT, or no disease (controls). Children either received TA or underwent observation. QoL questionnaires were administered at recruitment and 3 months later. Test-retest reliability was assessed using Bland-Altman plots. Preintervention scores were plotted against changes in scores, with pre-established cut-offs and cut-offs indicated by control group variability. Results: There were 120 children, 25 had no intervention, and 19 were controls. All questionnaires showed test-retest reliability over time. Using the distribution of scores from the control group we estimated the 95th percentile to redefine the cut-off for OSA-18 (reduced from 60 to 46) and PSQ (unchanged from 0.33). Higher pre-operative scores predicted greater reduction following TA, with OSA-18 the most consistent predictor of QoL change. The PSQ classified 86.8% of children undergoing TA above the 0.33 cut-off; whereas OSA-18 classified 73.7% above the 46 cut-off. Of these, 71.2% and 87.5% showed improvement after TA, respectively. Using the 95% confidence interval for change in the control group to identify a 'meaningful' change in score, children with OSA-18 scores >46 had a 93% chance of a meaningful improvement, whereas PSQ scores >0.33 were associated with an 80% chance of a meaningful improvement. Conclusions: OSA-18 is a better predictor of improved QoL than PSQ for TA in children with SDB. We propose a new cut off score (>46) for OSA-18. This may assist Otolaryngologists' decision making when assessing a child with SDB.