Leak rate doubles, AHI halves - what's up with that?

Awaiting Viewer 1.0 and experimenting a bit with pressures settings etc, about 2 weeks into therapy and logging the numbers on the machine daily (resetting to 0 of course) – thanks for teaching me about this fellow hosers. First off, my AHIs have been in the 4.5 to 5.5 range for a week and I’m clearly dreaming much much better so REM must be vastly improved. I’ve got a fairly specific REM problem I think (normal architecture, severe hypoxia restricted to REM). My alertness in the morning correlates quite well with the AHI values, even in this (what seems to me) rather narrow range. But I had a bit of a comfort issue with the medium OptiLife cushion at higher pressures 2 nights ago (not the “pillow” type that goes into the nostrel but the one that sits below and presses up, forget what it’s called ….), so I decided to switch to the small size.

I did feel more comfortable last night but before falling asleep I sensed that there was more of a tendency of the mask to leak with minor incorrect movements or bad positioning of the pillow, which makes sense I suppose (smaller contact area). I do prefer the smaller size in all other respects though. Sure enough the leak rate this morning was 45 L/min, compared to the usual 25-28 in the past week, though I did have 2 values in the 40s in the past week and my first week’s running average was 41. Maybe it’s not so bad. To my great surprise though, AHI was down to 2.6, i.e. about 50% of where it’s usually been! I do feel pretty good today, but not THAT much better than yesterday after an AHI 4.5 night. Nothing else has change that I can think of – same chinstrap (headband borrowed from my wife, fits great), same settings (min=7.0, max=16), no known feline interference during the night but you never know ….. 90% pressure was 8.0, pretty much in the normal range of 7 to 8.5 – this aspect seems rather stable. Pressure goes up to 10 or so when I’m in REM, based on one or two awakenings in the small hours at the end of a vivid dream. Nice (but a bit noisy for my wife unfortunately).

Any clever theories from the blower boffins out there? Is it too soon even to speculate? Maybe I need to wait for the full data picture which I shall soon have. If so, what should I be looking for? Last night’s AHI number does seem counterintuitive and I’m trained to try to try to interpret unexpected data as well as the predictable.

Also, is a leak rate of 45 something that the blower can compensate for fairly easily by increasing the flow rate? Does a higher leak rate much reduce the rate of response when I have a sudden event in REM? Intentional leak at 8 or 9 cm is about 25-28 with this set-up as I recall, it’s under 30 anyway. If I reduce the leak from 45 to the upper 20s with this small mask, can I expect even lower AHI values or am I getting greedy?

Thanks in advance.

As the leak rate increases, it becomes more difficult for an xPAP to detect events. Although your leak rate does not seem super high, it may be that your AHI is lower because your machine did not detect all of the event that occurred.

Thank you for that rapid response, which makes a good deal of sense. You would indeed expect the false negative rate to go up with increased blower compensation for leak.

Just the sort of thing I was hoping for. Man but this forum is an excellent resource ....