Echocardiographic screening in children with very severe obstructive sleep apnea.

OBJECTIVE: (i) To determine the prevalence of echocardiographic abnormalities in children with very severe OSA defined by an apnea hypopnea index (AHI) ≥ 30 events/hour. (ii) To test the hypothesis that polysomnographic parameters predict echocardiographic variables in this population.

METHODS: Children aged 1-17 years presenting with polysomnography demonstrating an AHI ≥30 and referred for pre-operative echocardiography performed within the 6 months prior to tonsillectomy and adenoidectomy (T&A), over a two-year period (January 1, 2016 to December 31, 2018) were evaluated. The exclusion criteria were the presence of (i) unrepaired congenital cardiac disease, (ii) tracheostomy, (iii) poorly controlled asthma, or (iv) neuromuscular disorder. The prevalence of echocardiographic abnormalities was determined for the study population. The impact of the severity of OSA on echocardiographic parameters was evaluated using Student's t-test. The relationships between polysomnographic variables and biventricular function as well as pulmonary hemodynamics were measured. A penalized regression model was used to identify the contributions of polysomnographic variables to each echocardiographic parameter by mitigating inter-variable relationships. P < .05 was considered significant.

RESULTS: Eighty-nine children were screened, of whom 47 were included for analysis. The mean age was 68.8 months [95% confidence interval, 56.0 to 81.6]. Thirty-three (70.2%) were boys. Twenty (42.6%) were obese. All children had normal echocardiograms. The differences in echocardiographic variables between children grouped by the severity of OSA were not statistically significant (P: 0.18-0.98). Polysomnographic variables predicted only 4 out of 13 studied echocardiographic parameters.

CONCLUSIONS: Pre-operative echocardiography did not identify significant abnormalities in children with very severe OSA. Majority of the echocardiographic variables were not predicted by polysomnographic parameters. This study demonstrates the limited benefit associated with routine echocardiographic screening of children with very severe OSA solely based on polysomnographic indices.