The present study was designed to assess whether isoflurane requirement was significantly affected by fresh gas flow in a closed-circuit system. Sixty patients scheduled for orthopaedic procedures were randomly assigned into three groups. In group A (n = 20), anaesthesia was conducted with a fresh gas flow of 482.5 +/- 186.6 ml X min-1, corresponding to the patient's metabolic demand. In group B (n = 20), the fresh gas flow was 2000 ml X min-1. In group C (n = 20), it was adjusted to the ventilation minute, i.e. 7145 +/- 986 ml X min-1. Artificial ventilation was conducted using a tidal volume of 10 ml X kg-1 and a rate of 10 to 12 c X min-1. Anaesthesia was induced after 10 min denitrogenation with fentanyl (4 micrograms X kg-1), thiopentone (4 mg X kg-1) and vecuronium (0.1 mg X kg-1). FIO2 was then brought to 0.5 in nitrous oxide and was monitored continuously using a polarographic oxymeter. Liquid isoflurane was injected in the expiratory limb of the circuit using an electrical syringe driver. Alveolar concentration of isoflurane was set at 0.92 vol. % according to Lowe and Ernst. Statistical analysis was carried out using Student's test for means. Anaesthesia lasted 138 +/- 88.3 min in group A, 125.5 +/- 45.1 min in group B and 146.5 +/- 50 min in group C, no difference being significant.(ABSTRACT TRUNCATED AT 250 WORDS)