OBJECTIVE: The objective of this study was to determine the extent to which adenotonsillar hypertrophy contributes to the severity of obstructive sleep apnea (OSA) in children. STUDY DESIGN: Thirty-three consecutive children who were referred to a sleep disorders center for evaluation of suspected OSA had standard lateral neck roentgenography performed. Adenoid size was determined by measuring the adenoidal-nasopharyngeal (AN) ratio. Tonsil size was quantitated on physical examination. The severity of OSA was determined by full-night polysomnography in the sleep laboratory. RESULTS: All of the patients reported snoring with trouble breathing, apneas, or both problems witnessed by a parent. The patients' respiratory disturbance index ranged from 0 to 95.3 (mean +/- SD 12.5 +/- 9.1). The patients' AN ratio ranged from 0.48 to 0.98 (0.76 +/- 0.14); 30 (91%) of the 33 patients had AN ratios greater than published normal means, and 16 (48%) had AN ratios more than 2 standard deviations above published means. Although the AN ratio and tonsil size did not predict the number of apneas, a significant relationship was seen between the AN ratio and the duration of obstructive apneas (r = 0.48, p < 0.01). Obesity (percent ideal body weight) was the only independent predictor for the number of respiratory events per hour of sleep (r = 0.49, p < 0.01). Percent ideal body weight was also the major predictor of the lowest oxyhemoglobin saturation (r = -0.58, p < 0.0001), but the AN ratio also contributed to the variance in saturation, with a correlation coefficient (r) of 0.69 for the two factors (p < 0.0001). CONCLUSION: Lymphoid hyperplasia affects the severity of apnea more than the number of obstructive apneas. The pathophysiologic characteristics of OSA in children probably involve complex interactions between pharyngeal size and mechanics.