We performed low-flow sevoflurane anaesthesia at a flow rate of 1 litre min-1 in three groups (n = 8 each) using 600 g of fresh soda lime (control group), 600 g of soda lime with 60 ml of water added (water group) or 600 g of soda lime saturated with carbon dioxide, that is partly exhausted soda lime (carbon dioxide group). Degradation products in the system were measured hourly. Inspired and end-tidal carbon dioxide and sevoflurane concentrations, carbon dioxide and temperature of the soda lime were monitored. CF2 = C(CF3)-O-CH2F (compound A) was the only sevoflurane degradation product detected. The mean maximum concentration of compound A was significantly higher in the control group (mean 16.0 (SD 5.0) ppm) than in the water (1.4 (1.0) ppm) or carbon dioxide (4.0 (1.8) ppm) group, and the maximum temperature of the soda lime was significantly lower in the carbon dioxide group (30.7 (3.5) degrees C) than in the control (43.4 (1.8) degrees C) or water (40.8 (1.8) degrees C) group (P < 0.05). The use of partly exhausted soda lime or soda lime with water added reduced compound A concentrations in the system during low-flow sevoflurane anaesthesia.