We have compared the incidence of preoperative and postoperative oxygen desaturation in 20 elderly patients undergoing dynamic hip screw insertion for fractured neck of femur, allocated randomly to two groups to receive subarachnoid anaesthesia (SA, n = 10) or a general anaesthetic (GA, n = 10). Oximetry data were recorded during the preoperative night and the first 48 hours after surgery using Satmaster. Data associated with zero amplitude signal were automatically invalidated by the software and decreases in SpO2 which were preceded by contemporaneous changes in signal amplitude which conformed to a previously described template were assumed to be artefactual and were discarded from final data analysis. Data demonstrated a wide interpatient variability. However, those patients who desaturated preoperatively continued to do so in the postoperative period. The differences between the preoperative and postoperative oximetry profiles were examined for each patient and demonstrated a significantly longer time spent with SpO2 < 90% in the GA group compared with the SA group, in all the recorded time periods, except on the day of surgery. The SA group showed an improvement in oxygen saturation postoperatively when compared to the preoperative night, spending less time with an SpO2 < 85%. We conclude that the subarachnoid anaesthetic technique was associated with a lower incidence of postoperative oximetry desaturation when compared with general anaesthetic for these elderly patients undergoing repair of fractured neck of femur.