Leak Rate Questions

Hello All! For my first post, I have some questions about leak rates. I recently got a ResMed S9 AutoSet, and a bit later I installed ResScan and SleepyHead on my computer. I have now accumulated 9 days of data, and the leak rates are not what I would have expected. Although my AirFit P10 mask doesn’t vent as much air as I think it should to prevent rebreathing significant amount of the previous exhalation, it definitely does continuously vent at least some air. However, the readings I get from the software make it look like the leak rate is zero, other than some occasional spikes when my nasal pillows get dislodged. For example, on the statistics tab of ResScan, the median and 95th percentile leak rates show 0.0 for all 9 days. The maximums by day were 0.0, 1.2, 0.0, 15.6, 14.4, 2.4, 2.4, 1.2, and 0.0. On the ResScan detailed graphs, the leak rate is zero almost all the time on each of the days, with a few spikes here and there of almost no time duration.

The results are similar with SleepyHead, although the numbers differ. Under Daily Details, my minimum, median, and 95% readings are 0.00 for all 9 days. The maximums by day were 3.60, 4.80, 3.60, 60.00, 6.00, 15.60, 7.20, 4.80, and 2.40. On the Sleepyhead daily graphs, the graphs show zero (or a few teensy spikes) just about all the time, with the exception of the 60.00 day and the 15.60 day, which showed a few clearly significant but extremely brief spikes.

Is leak rate supposed to be an absolute or a relative measure? If it should be absolute, I would say that something is wrong with the measurement system in my S9. If it is relative, does it have a baseline amount of leakage that it subtracts from all readings that corresponds to the fact that the S9 settings have nasal pillows as the selected mask type? And even if that is the case, why are the daily maximums shown in ResScan so different than those reported by SleepyHead? (The two programs don’t even agree on which day had the second highest leak rate.) Thanks in advance for your help!

The leak rate reported by ResMed equipment is the net unintentional leak rate.
Each mask has a built-in leak rate that has to do with the removal of the excess CO2 from your exhalations.
If you take a look in the user manual that came with your P10 mask, you can see the expected leak rate for various pressures shown on page 7. Add that to the reported rate and you get the actual total leak rate.

If your mask had no excess leaks, i.e., perfect fit, then the reported rate would be 0.0.

MotorBoater wrote: Is leak rate supposed to be an absolute or a relative measure? If it should be absolute, I would say that something is wrong with the measurement system in my S9. If it is relative, does it have a baseline amount of leakage that it subtracts from all readings that corresponds to the fact that the S9 settings have nasal pillows as the selected mask type? And even if that is the case, why are the daily maximums shown in ResScan so different than those reported by SleepyHead? (The two programs don’t even agree on which day had the second highest leak rate.) Thanks in advance for your help!

your autoset is working as designed, and reporting excessive leaks.

most all masks leak at around the same amount, give or take a few lpm. the machine knows, in general, what masks are, and can then calculate a base actual leak rate by observing the flow as you breath... (lots of math involved, explained in another post somewhere on the forum here).

if it weren't venting co2, you'd be dead by morning. (well, you'd be dead a lot faster than morning, but you get the picture.)

I *have* heard reports on another board of a user who's p10 vents got clogged (he said they looked different, more plasticy or something) and were not venting at all, it was easy to tell that there was no venting, both by the stale air rebreathed, and not feeling the normal gentle airflow. he said his mask vendor also told him they'd gotten some reports on the issue.

scrubbing the vent mesh with a toothbrush under hot water restored normal operation.

note that scrubbing the vent mesh is part of the normal cleaning procedure resmed says to do.

Thanks JDS74 and palerider! I had a funny feeling that the reported leak rate was the amount by which the absolute leak rate exceeds the expected leak rate, based on the chosen mask type. From your responses, that appears to be the case.

I wish that I could also see the absolute leak rate, given that the mesh on my AirFit P10 is so restrictive and could clog rather easily without my knowledge, due to the tiny mesh holes. I have been scrubbing the mesh once a week with a brush; I hope that is enough. If I could measure absolute leak rate, I would be able to find out each morning whether the mesh is still behaving within specifications. That would alert me as to whether the mesh would need a more thorough cleaning or even replacement. Since my system does not report a negative relative leak rate if the leakage is below the expected value, I guess there is no way for the machine or software to tell me if the mesh is operating as intended. Darn!

Using the same set up, seeing the same net leakage type results and getting 8 hrs sleep with AHIs of 1, so I am happy to support that ResMed measures "impactful" leakage. The P10 is very odd in the way it "leaks" constantly but so darned comfortable.

MotorBoater wrote:Thanks JDS74 and palerider! I had a funny feeling that the reported leak rate was the amount by which the absolute leak rate exceeds the expected leak rate, based on the chosen mask type. From your responses, that appears to be the case.

I wish that I could also see the absolute leak rate, given that the mesh on my AirFit P10 is so restrictive and could clog rather easily without my knowledge, due to the tiny mesh holes. I have been scrubbing the mesh once a week with a brush; I hope that is enough. If I could measure absolute leak rate, I would be able to find out each morning whether the mesh is still behaving within specifications. That would alert me as to whether the mesh would need a more thorough cleaning or even replacement. Since my system does not report a negative relative leak rate if the leakage is below the expected value, I guess there is no way for the machine or software to tell me if the mesh is operating as intended. Darn!

the PR machines report total leak, the resmeds report excess leak.

I just feel for the gentle breeze from the p10 and figure as long as I feel that, it's probably good enough... though it would be nice to know if it was getting below it's base 20lpm@4cm. hard to believe that it's venting near 50lpm at 20cm... I get up in that neighborhood at times and it really doesn't feel like it... but, then, it doesn't feel like it on the pilairo q either, if it's diffuser is clipped on.

is it because the diffusers on the P10 are so big that they allow high flow but don't impact pressure much.

Hello again! I just got back from vacation and saw that there had been more posts on this thread during that time. One thing that nobody had addressed yet is the significant difference in the reported maximum leak rate numbers (and which days were worst) between ResScan and SleepyHead. Does anyone reading this thread know why these two programs would differ in such a way, or should I post that part of the question again as a new topic with a different title?

I don't know why the maximum leak values differ between the 2 software programs but I have seen that happen myself.
Sometimes they match and sometimes they don't. They are always close though. I might see a spike even on ResScan to say 20 L/min yet ResScan says my maximum was only 15 L/min. SleepyHead will show the 20 but ResScan doesn't.
Wonder if Mark can explain it? He's the guy who came up with SleepyHead.
Never worried me enough to try to sort it out so I can't give you a definite answer as to why...only that I have seen it myself.

Pugsy wrote:I don't know why the maximum leak values differ between the 2 software programs but I have seen that happen myself.
Sometimes they match and sometimes they don't. They are always close though. I might see a spike even on ResScan to say 20 L/min yet ResScan says my maximum was only 15 L/min. SleepyHead will show the 20 but ResScan doesn't.
Wonder if Mark can explain it? He's the guy who came up with SleepyHead.
Never worried me enough to try to sort it out so I can't give you a definite answer as to why...only that I have seen it myself.

it's possible that resscan is ignoring, or smoothing over brief spikes.

I'll ask mark, though he's busy fixing an annoying bug I just found... gotta keep those programmers busy, else they'll kern your fonts or round the corners on your pop up boxes...

There is a setting in SleepyHead preferences that allows you to choose between the true maximum value and 99th percentile to show as the maximum, which filters out cruft spikes from messing up your maximum values..

If you like your spikes spikey, set it to true maximum.. (Personally, I don't like spikey data, I like my numbers to mean something, and prefer random meaningless spikes that happened for a brief few seconds neutralised)

An earlier version (prior to 0.9.6 ) had these values accidentally switched, so invert accordingly if you must..

Edit: I'm not sure if ResScan cruft filters by default... I don't run windows, and I've never even seen the program apart from occasional screenshots.. I can tell you if they do cruft filter, it's highly probable their cruft filter calculations won't match mine exactly, unless by chance they chose the same method and percentile value. I'm not out to clone ResScan, and both methods I provided are statistically valid.

Thanks Pugsy, palerider, and especially jedimark (!) for clarifying that it is not abnormal for the two programs to differ with regard to maximum leak readings.