Autologous transfusion in total knee replacement surgery.

EditorÐRahman and colleagues 1 make the interesting observation that obstructed breathing in post-operative patients was more common when they were awake than when they were asleep. This seems to contradict our study 2 where we found obstructed breathing postoperatively only when patients were asleep. There were important differences between the two studies. We 2 divided 32 patients into two equal groups, one having morphine, the other regional block. The dose of morphine was about 28 mg in 14 h and no opioid was used in the regional block group. In Rahman and colleagues' study three patients had a median dose of 51 mg morphine in the 14 h study period and ®ve had thoracic epidural blocks which included opioids. There are differences in the criteria of obstruction. Rahman and colleagues 1 de®ned obstruction as a phasic abdominal EMG pattern at any time or any change in nasal ¯ow pattern during a 30 s measuring period. My reading of this is that one obstructed breath in a 30 s measuring period registered as a 30 s period of airway obstruction and if every breath was obstructed during the 30 s period it would give the same duration of obstruction. Our criteria and methods were different, 2,3 paradoxical breathing (partial obstruction) was where 50% of breaths in a 5 min period showed phase difference between the movement of ribs and abdomen and obstructive apnoea was recognized only when the sum of rib cage and abdominal movement contributing to tidal volume was zero for more than 10 s. This means that Rahman and colleagues 1 would have detected many, very brief (< 10 s), periods of apnoea inàwake' patients all of which we would have ignored. In our study, partial obstruction with snoring 3 lasting 5 or more min was the most common breathing pattern associated with hypoxaemia (SaO 2 < 80%) but occurred only during sleep and only after morphine. A further crucial difference is the assessment of thèsleep/ awake' state. While both studies used the same EEG criteria, Rahman and colleagues 1 de®ned patients asàwake' up to EEG stage 2 whereas we de®ned a `transition stage' of sleep at EEG stages between 1 and 2. Thus we labelled asàsleep' some patients that Rahman and colleagues 1 would have labelledàwake' using their criteria. Nevertheless, in this transitional group we found only 94 episodes of obstruction compared to 510 episodes when asleep. In …