Previous investigators have demonstrated variable responses to uvulopalatopharyngoplasty (UPP) in patients with obstructive sleep apnea. We hypothesized that this variability is due to either (1) differences in baseline pharyngeal collapsibility preoperatively or (2) differences in magnitude of the decrease in pharyngeal collapsibility resulting from surgery. To determine the relationship between changes in collapsibility and the response to UPP surgery, we measured the upper airway critical pressure (Pcrit) before and after UPP in 13 patients with obstructive sleep apnea. During non-REM sleep, maximal inspiratory airflow (VImax) was quantitated by varying the level of nasal pressure (PN), and Pcrit was determined by the level of PN below which VImax ceased. A positive response to UPP was defined by a greater than or equal to 50% fall in non-REM disordered breathing rate (DBR). In the entire group, UPP resulted in significant decreases in DBR from 71.1 +/- 22.4 to 44.7 +/- 38.4 episodes/h (p = 0.025) and in Pcrit from 0.2 +/- 2.4 to -3.1 +/- 5.4 cm H2O (p = 0.016). Moreover, the percent change in DBR was correlated significantly with the change in Pcrit (p = 0.001). Subgroup analysis of responders and nonresponders demonstrated that significant differences in Pcrit were confined to the responders. Specifically, responders demonstrated a significant fall in Pcrit from -0.8 +/- 3.0 to -7.3 +/- 4.9 cm H2O (p = 0.01), whereas no significant change in Pcrit was detected in the nonresponders (1.1 +/- 1.6 versus 0.6 +/- 2.0 cm H2O. No clinical, polysomnographic, or physiologic predictors of a favorable response were found preoperatively.(ABSTRACT TRUNCATED AT 250 WORDS)