Background Previous studies in humans suggest that inhibition of upper airway muscle activity is independent of the dose of inhalational anesthesia. Whether a dose-independent relation applies to changes in airway caliber is unknown. The authors sought to evaluate the configurational changes that lead to upper airway narrowing during inhalational anesthesia with sevoflurane and to determine whether these changes are dose dependent within a clinically relevant dose range. Methods Fifteen children undergoing elective magnetic resonance imaging of the brain were studied. Magnetic resonance images of the upper airway were acquired at sevoflurane concentrations of 0.5, 1.0, and 1.5 minimum alveolar concentration (MAC), administered in random sequence. At least 15 min was allowed for equilibration of inspired and alveolar partial pressures of sevoflurane. Images were acquired in early expiration at the level of the soft palate, base of the tongue, and tip of the epiglottis. Airway cross-sectional area (CSA), anteroposterior, and transverse dimension were determined using image-analysis software. Results At each anatomical level, pharyngeal CSA decreased progressively with increasing depth of sevoflurane anesthesia (P < 0.001). Increasing the sevoflurane concentration from 0.5 to 1.0 MAC reduced airway CSA by 13-18%, and a further increase to 1.5 MAC resulted in an overall 28-34% reduction in CSA. The reduction in CSA was predominantly due to a decrease in anteroposterior dimension. Conclusions Increasing the depth of sevoflurane anesthesia resulted in a relatively uniform reduction in pharyngeal caliber at each anatomical level studied. The effect of sevoflurane on upper airway caliber is dose dependent.