B80-3-12 Hypoxic Burden and Blood Pressure Trajectory in Pediatric Obstructive Sleep Apnea: A Secondary Analysis of the Chat Clinical Trial

Rationale

Hypoxic burden (HB) quantifies cumulative hypoxemia severity in obstructive sleep apnea (OSA) and predicts cardiovascular outcomes in adults. Whether HB is associated with blood pressure (BP) trajectory in children with OSA, and whether adenotonsillectomy (AT) modifies this relationship, remains unknown.

Objectives

To examine associations between baseline HB and BP trajectory in children with OSA, test whether AT modifies this relationship, and assess whether HB reduction is associated with BP improvement.

Methods

Secondary analysis of the Childhood Adenotonsillectomy Trial (CHAT). HB (%·min/h) was calculated from baseline polysomnography and z-standardized. The primary outcome was change in systolic BP percentile (ΔSBP) from baseline to 7-month follow-up; change in diastolic BP percentile (ΔDBP) was secondary. Linear regression examined baseline HB and BP associations: Model 1 (unadjusted) and Model 2 (demographics: age, sex, BMI z-score, Black race). Treatment effect modification was tested via HB×AT interaction. Mechanistic analyses examined whether AT reduced HB and whether HB reduction was associated with BP improvement using correlation and tertile comparisons.

Results

Among 383 children (age 6.6±1.4 yr; 49% male; BMI z-score 0.9±1.3; 54% Black; AHI 6.7±5.6 events/h), mean HB was 7.8±8.1 %·min/h. Baseline HB was associated with ΔSBP (M2: β = 3.73; 95% CI, 0.64, 6.82; p = 0.02). Stratified analyses showed consistent effect sizes across both treatment arms (AT: β = 3.42; 95% CI, -1.31, 8.16; p = 0.16; watchful waiting [WW]: β = 3.98; 95% CI, -0.15, 8.11; p = 0.06). When the HB×AT interaction term was introduced, the main effect of HB remained similar in magnitude (β = 3.86; 95% CI, -0.58, 8.30; p = 0.09), with no significant interaction (p = 0.92). No associations were observed with ΔDBP. AT reduced HB by 72% compared with WW (p < 0.001), yet AT was not associated with ΔSBP (β = 1.32; 95% CI, -4.68, 7.32; p = 0.67). Change in HB showed minimal correlation with ΔSBP (r =-0.01; p = 0.90), though tertile analysis comparing T3 (greatest HB reduction) versus T1+T2 (minimal and moderate) suggested a modest trend toward greater SBP improvement (β =-3.84; 95% CI, -10.53, 2.86; p = 0.26).

Conclusions

Children in CHAT had low overall HB. Higher baseline HB may be associated with worsening SBP trajectory, with consistent effect sizes but wider confidence intervals in stratified and interaction models. Despite substantial HB reduction following AT, surgery was not associated with short-term BP improvement. HB may serve as a cardiovascular risk marker, though the low baseline HB and short follow-up may have limited our ability to detect effects. Longer-term studies in children with greater hypoxemic burden are needed.

This abstract is funded by: T32 HL007901