Pediatric Obstructive Sleep Apnea: Predictors of Surgical Success via Weighted Analysis

Objective To identify predictors of surgical success after pediatric tonsillectomy for obstructive sleep apnea (OSA) and estimate the degree to which non‐random selection into postoperative polysomnography (PSG) distorts outcome associations. Methods Retrospective nested cohort study at a quaternary pediatric center (2016–2024). From 940 children undergoing tonsillectomy for OSA, 389 had preoperative PSG and 174 had both preoperative and postoperative PSG. Inverse probability weighting (IPW) corrected for non‐random selection into follow‐up PSG. Surgical success was defined as postoperative AHI < 5 events/h. Results The outcome cohort included 174 children (median age 6.0 years; 62% male; 45% obese; median preoperative AHI 16.5 events/h). Surgical success was achieved in 122 (70.1%). Follow‐up PSG was associated with higher preoperative AHI (OR 1.06 per event/h), medical complexity (OR 2.22), and multiple comorbidities (OR 1.86), with good discrimination (AUROC 0.795). In IPW‐adjusted analysis, each additional event/h of preoperative AHI was associated with lower odds of surgical success (OR 0.96 per event/h; p < 0.001). The estimated effect of obesity strengthened after weighting (OR 0.53 vs. 0.66 unweighted), though the association did not reach statistical significance ( p = 0.15). Children with very severe OSA (AHI > 30) achieved success in only 41.7%. Conclusion Higher preoperative AHI was associated with lower odds of surgical success after tonsillectomy. The obesity estimate did not reach statistical significance after weighting, but whether this reflects a true absence of effect, limited power, or confounding by AHI severity cannot be determined from this dataset. Children with very severe OSA should be counseled that surgery alone is unlikely curative. Level of Evidence 4.