Hyperoxaemia depresses the output of peripheral and central chemoreceptors. Patients treated with opioids often receive supplemental oxygen to avert possible decreases in oxygen saturation (Sp(O2)).We examined the effect of a single dose of remifentanil in healthy volunteers inhaling room air vs air enriched with 50% oxygen. Twenty healthy volunteers received i.v. 50 mg remifentanil (infused over 60 s) at anormoxic (N) or hyperoxic (FI(O2) 0.5, H) background on separate occasions. Minute ventilation (Vi), respiratory rate (RR), end-tidal PC(O2), and Sp(O2) were collected on a breath to-breath basis. The occurrence of apnoea was recorded. During normoxia, remifentanil decreased Vi from 7.4 (1.3) [mean (SD)] to 2.2 (1.2) litre min 21 (P,0.01), and during hyperoxia from 7.9 (1.0) to 1.2 (1.2) litre min 21 (P,0.01; H vs N: P,0.001). RR decreased from 13.1 (2.9) to 6.1 (2.8) bpm during N (P,0.01) and from 13.2 (3.0) to 3.6 (4.0) bpm during H (P,0.01; H vs N: P,0.01). During normoxia, Sp(O2) decreased from 98.4 (1.5) to 88.6 (6.7)% (P,0.01), while during hyperoxia, Sp(O2) changed from 99.7 (0.7) to 98.7 (1.0)% (P,0.001). Apnoea developed in two subjects during normoxia and 10 during hyperoxia. Respiratory depression from remifentanil is more pronounced in hyperoxia than normoxia as determined from minute ventilation, end-tidal PC(O2), and RR. During hyperoxia, respiratory depression may be masked when measuring Sp(O2) as pulse oximetry remains in normal values during the first minutes of respiratory depression.