Leak = no events???

[jbn3boys]

Okay, so this makes NO sense to me. When I looked at my information from last night, I had nearly an hour when I had nearly no events. That's HUGE for me! Previously, I was doing good to get 15 minutes with no events. But then I looked at my leak rate, and during that exact hour, my leak rate was much higher than normal. So, if I have a higher leak rate, that means fewer events? That seems backwards to me.

I'm currently on bipap with pressures of 15/11 (second night at that level) The leak rate for my mask at 14 is 36. I'm hoping to head in to the DME today to pick up a new set of pillows.

[Marvinvwinkle]

I don't think your leak rate was above what is considered normal. That hour almost looks like the machine stop recording for an hour, but it did show a leak rate that was a little above average. I would be more concerned with the AHI, and the number of events that are taking place of which I sure you are aware.

[Pugsy]

The leak isn't all that huge and well within the ability of the machine to handle.

Odd that no events during that time but could just be a coincidence unless you see this pattern frequently. I know your situation is unique and troubling for you. I am sorry that I can't offer any new ideas for you.

[robysue]

jbn3boys,

I still cannot figure out how the System One BiPAPs figure out when to label a leak as a "large leak."

The Resmed S9 documentation was much clearer about what constituted a leak that was large enough to potentially interfere with the accuracy of the data and the quality of the therapy: On the S9, long leaks at an unintentional rate of 24 L/min or greater are considered large enough to potentially interfere with the machine's ability to correctly detect events, respond to events (for the S9 AutoSet), and deliver effective therapy.

If we apply this Resmed definition of "large leak" (instead of the unknown PR definition), then that hour-long leak is right underneath the point at where unintentional leaks become problematic: 36+24 = 60 L/min and it looks to me like you were leaking at a rate between 50 and 60 L/min during that hour.

[Pugsy]

Was there a corresponding increase in pressure during that hour in question? You are using FF mask? Yes? At your pressure what is the expected vent rate? If variable pressure the vent rate for the "average"?

Robysue.. I have never been able to find out exactly what Respironics flags as "large leak" either but from my past reports it seems to be flagged around 100 L/min. I have seen this on reports when I wake up and mask has become unseated in one nostril. A "minor" unseating might show me 80 L/min and no Large Leak flag. I think that the Respironics machines historically seem to be a bit weird with their leak reporting. Since we can't dial in a definite vent rate for a particular mask choice it becomes not as concrete as the ResMed line for large leak. Would be nice if it was but it just happens to be a close estimate at best. Since you are a math person I can just imagine how that bugs you but they just don't give us enough concrete data to make concrete statements.

If jbn had a pressure increase at the time of the increased leak we might wonder if the machine dealt with events with the increase in pressure and thus created the chance for the leak. It could just be a quirk. That is why I mentioned if this seemed to occur and establish a pattern. I don't think the leak was large enough to invalidate the data or lack of events at least from this one occurrence.

[jbn3boys]

... I would be more concerned with the AHI, and the number of events that are taking place of which I sure you are aware.

I am continually perplexed by the high AHI. I can NOT seem to get that under control, even after six months of therapy at 100% compliance. I am currently working with my sleep doc, but it is a very long, very slow process, complicated by the fact that during my sleep study last week, I had only three recorded respiratory arousals the whole night. (Still can't figure that one out!)

robysue, in EncorePro, you can change to showing "unintentional leak rate" in the company/preferences. If I do that, then the average leak rate shows up as 2.8 (I think), and it looks like that larger leak is indeed between 20 and 25 L/m. However, I don't know how EncorePro determines what is "unintentional", since I don't think I ever put in what mask I am using. My guess would be that it uses the "average" (mode) leak rate to be the intended?

Was there a corresponding increase in pressure during that hour in question? You are using FF mask? Yes? At your pressure what is the expected vent rate? If variable pressure the vent rate for the "average"?

I am on straight biPAP, so there is no increase in pressure, only the "puffs" of air to determine if an event is obstructive or not, or to stimulate a breath. I am using Swift FX, but I don't have much problem with mouth breathing. I would say "no" mouth breathing, but during my sleep study the tech told me that three times during the night I did open my mouth, but only a little puff of air and then closed again. My husband knows to listen for mouth breathing, and has never said anything about it. I listed the intended leak rate as 36 at a pressure of 14 (my pressure is 15/12)

As far as a pattern, it seems to me that I have noticed before when there is an increased level of leak that there is a corresponding "break" in events, but it has never been this long. As I said, I am usually lucky to get 15 minutes at a time with no events. I will go back and look more closely at leaks that I've had recently. I'm just so tired of not getting good, restful, restorative sleep.

[Pugsy]

owever, I don't know how EncorePro determines what is "unintentional", since I don't think I ever put in what mask I am using. My guess would be that it uses the "average" (mode) leak rate to be the intended?

This is what I have come up with also. When I play with total leak vs unintentional it appears to be approx 28 for my swift FX. I just compared the 2 on last nights reports. With total leak I get a 34 L/min and with unintentional leak I get 6.4. How it comes up with the 28 without knowing mask has to just be a close estimate I guess. I sometimes will hit 50 and 60 L/min on the total leak without any increase in pressure (I am on APAP) and sometimes have events scored and sometimes no events. Like your little increase they might last 30 minutes to an hour. I can't see much of a pattern with mine. Sometimes a single event, sometimes nothing with no increase in pressure. When I have a leak with a pressure increase I don't pay a whole lot of attention to it because I don't know who came first...

[jbn3boys]

I just looked back at the last week or so of data. It appears quite often that if I have a larger leak, that it will correspond with a lack of events. My events have always been scattered throughout the night, so when there is a gap in events, it is pretty obvious. I don't know how long this has been going on. I guess I'll have to do some more digging to find out.

I'm also waiting on word from my doc about my iron/ferretin levels, to see if I might also have PLMD/RLS. That could be another factor in my numbers not coming down to where they are.

But I still don't understand how I could have 3 events all night at the sleep lab, and 7.9 per HOUR last night. Makes no sense to me. (And actually, if you take out that hour with no events, my AHI would be even higher.)

[robysue]

robysue, in EncorePro, you can change to showing "unintentional leak rate" in the company/preferences. If I do that, then the average leak rate shows up as 2.8 (I think), and it looks like that larger leak is indeed between 20 and 25 L/m. However, I don't know how EncorePro determines what is "unintentional", since I don't think I ever put in what mask I am using. My guess would be that it uses the "average" (mode) leak rate to be the intended?

I know that EncorePro lets you change the "total leak" to "unintentional leak" but I've left it at "total leak" in my copy of EncorePro exactly because I can't seem to determine how EncorePro decides what value to subtract off since there seems to be no way of telling the program what mask your using---or even what type of mask your using. Also, for reasons that I don't understand, my own total leak rates tend to run a bit under the advertised expected leak rate of the Swift FX mask. As a mathematician, I assume the graph in the FX manual gives approximate or average expected leak rates, but unlike other mask manuals, it doesn't give an error range---i.e. other manuals often say something along the lines of

• "expected leak rate at x cm of pressure is y L/min plus or minus 1.5 L/min"

for example.

But in any case, I do think jbn3boys's unintentional leak rate for the hour in question is between 20 and 25 L/min, which is right around Resmed's RedLine in ResScan. Does that mean anything on a PR System One? I don't know. But since there seems to be no available information about when the unintentional leak rate becomes problematic on the PR System One, seems to me like a reasonable guessimate is that it's in the same ball park as the Resmed S9. In that case, the data might indeed be somewhat compromised. Particularly since jbn3boys goes on to say (in response to a question from Pugsy):

As far as a pattern, it seems to me that I have noticed before when there is an increased level of leak that there is a corresponding "break" in events, but it has never been this long. As I said, I am usually lucky to get 15 minutes at a time with no events. I will go back and look more closely at leaks that I've had recently. I'm just so tired of not getting good, restful, restorative sleep.

and

I just looked back at the last week or so of data. It appears quite often that if I have a larger leak, that it will correspond with a lack of events. My events have always been scattered throughout the night, so when there is a gap in events, it is pretty obvious. I don't know how long this has been going on. I guess I'll have to do some more digging to find out.

And concerning the more general issue of large leaks on the PR System One, Pugsy writes:

Robysue.. I have never been able to find out exactly what Respironics flags as "large leak" either but from my past reports it seems to be flagged around 100 L/min.

No wonder that even my largest leaks ever don't get labeled as a Large Leak: At the pressure range I'm now using (7/4), the expected leak rate for the Swift FX is 20L/min (at 4cm) and 25 or 26 L/min (at 7cm) and my daily data for ever since the pressure reduction runs between 18 L/min to 20 L/min with nice, flat leak lines. So the PR System One won't flag a leak until my leak rate is roughly four or five times as large as my expected leak rate? As a user with prescribed low pressure settings, I find that somewhat alarming.

I believe that the quality of my therapy and the reliability of the data would be seriously compromised well before I had a long leak that's four of five times the expected leak rate of the mask. Using the Resmed algorithm for labeling something as a Large Leak, my therapy and data would start to become questionable at roughly twice my expected leak rate. On the one night since switching to the S1 BiPAP where I did have some 15 minute long leaks that high, however, the PR System One recorded no Large Leaks for the night. That night was the night I had a monster migraine and I had the headgear for the FX adjusted ridiculously loose because the normal "very loose" adjustment that I typically use was still enough to aggravate the migraine.

I have seen this on reports when I wake up and mask has become unseated in one nostril. A "minor" unseating might show me 80 L/min and no Large Leak flag. I think that the Respironics machines historically seem to be a bit weird with their leak reporting. Since we can't dial in a definite vent rate for a particular mask choice it becomes not as concrete as the ResMed line for large leak. Would be nice if it was but it just happens to be a close estimate at best. Since you are a math person I can just imagine how that bugs you but they just don't give us enough concrete data to make concrete statements.

As a math person, I'm concerned that the Large Leak flag in the Resprionics machine is NOT a close estimate at all. I'm also concerned that there seems to be NO documentation in (a) the user manual for the machine itself, (b) the documentation for EncoreViewer, and (c) the documentation for EncorePro that lets a person determine just what the PR engineers think a Large Leak is and when the PR engineers believe the quality of the therapy and the reliability of the data starts to get compromised. Resmed is acting much more responsible here---even if their choice of Mr. Green Smiley Face and Mr. Red Frowny Face are insultingly stupid ways to tell the user there is a serious problem with leaks.

[Pugsy]

As a math person, I'm concerned that the Large Leak flag in the Resprionics machine is NOT a close estimate at all. I'm also concerned that there seems to be NO documentation in (a) the user manual for the machine itself, (b) the documentation for EncoreViewer, and (c) the documentation for EncorePro that lets a person determine just what the PR engineers think a Large Leak is and when the PR engineers believe the quality of the therapy and the reliability of the data starts to get compromised. Resmed is acting much more responsible here---even if their choice of Mr. Green Smiley Face and Mr. Red Frowny Face are insultingly stupid ways to tell the user there is a serious problem with leaks

I hear you and I agree but not much we can do about it. Frustrating? Yes. I just have to tell myself that Respironics doesn't really cater to me (the end user) and the DME's and doctors that they do cater to, don't/won't look at individual items. They just look for general "trends" if they look at all. Does it make it right? Of course not but other than getting a ResMed machine, nothing I can't do about it. Since I rarely have much variation with my reports anymore, it doesn't bug me so much but I can sure see why it bugs jbn. That is why I asked if she could see a pattern and she can, so it may warrant further investigation.

I am curious, if you toggle between Total Leak and Unintentional leak on your reports, what is the difference?
Are they using the same roughly 25 cm that I get or is your a different number since your pressure is lower than my average of 11 cm?

[jbn3boys]

Since I rarely have much variation with my reports anymore, it doesn't bug me so much but I can sure see why it bugs jbn.

I think I'll just send you two all my data from the last two months and see what YOU can come up with!

[Pugsy]

I think I'll just send you two all my data from the last two months and see what YOU can come up with!

Oh man, from just what I have seen lately I might be cross eyed really quickly. I can certainly understand your frustrations and looking under every rock for an answer. I would be doing the same thing. I sure wish I could think of something that hasn't been thought of before but my bag of ideas is also depleted.

I was lucky, plain old vanilla OSA worse in REM, easy to treat once we got the pressures right. My other physical issues kept me from seeing dramatic results so I had to accept the fact that I was "preventing" worse stuff. I had nothing that compares to your situation.

Better send them to Robysue, she is much more of a stickler for math details than I am.

[DreamDiver]

As a math person, I'm concerned that the Large Leak flag in the Resprionics machine is NOT a close estimate at all. I'm also concerned that there seems to be NO documentation in (a) the user manual for the machine itself, (b) the documentation for EncoreViewer, and (c) the documentation for EncorePro that lets a person determine just what the PR engineers think a Large Leak is and when the PR engineers believe the quality of the therapy and the reliability of the data starts to get compromised. Resmed is acting much more responsible here---even if their choice of Mr. Green Smiley Face and Mr. Red Frowny Face are insultingly stupid ways to tell the user there is a serious problem with leaks.

When I was using the M-Series Pro, this is how I figured out large leak:

Find out what the intended leak rate is for the mask you have at the pressure you're set for from the paperwork that came with it. For instance, for the mirage quattro full face go to page 11 of the manual. It suggests:

At 8cmH20 its 32 L/min +- 6L/min.
At 12 its 41.
At 16 its 48.
At 20 its 54.

For a pressure of 11, I can see from the graph on page 11 that my intended leak rate should be about 38 L/min. So long is the leak remains relatively steady at about 35 to 50 Liters/min on the total leak graph, I don't worry too much about leak. If it gets to around 70 - 80 L/min, I would start to get concerned whether I'm actually receiving therapy.

For the SwiftFX for her, at a pressure of 14, the flow curve on page 5 once again brings us to about 40 L/min. Because it's likely that the 50L/min episode was caused by mouth leakage, perhaps event detection is partially hampered even though 'relatively' efficient therapy is being delivered. I hope others will weigh in with better or more useful information.

[HoseCrusher]

Perhaps it would be a good idea to back up a little and gain some perspective on the larger picture.

IF a machine has a maximum airflow of 200 liters per minute, and IF an average value for bypass leak for a mask is 40 liters per minute at average pressures used, then a large leak alarm set for 150 liters per minute seems to be a good value. This gives the system a margin of error of 10 liters per minute.

All of this is just looking at the machines ability to deliver pressure in the presence of leaks.

Now, it could be that the resolution of the machine scoring is much better when the leak rate is equal to the bypass leak rate of the mask.

If this is close to being correct, then the machine is capable of delivering therapy at leak rates up to the large leak rate, but the scoring algorithms during higher leak rates may not provide good information.

An interesting way to check this out would be to poke a small hole in your hose about 1 foot from the machine and in an area that is covered by your hose cover. Sleep with this "extra" leak for a few nights and see if the data changes. It may take more than one hole, and you could probably look at your mask vent size and go with that. That would be like having a leak equal to twice your normal mask leak rate. If you could find the proper plumbing fixtures, you could put a Tee fitting in at the machine and put a valve in to allow you to adjust the leak rate while still being able to use your mask.

[jbn3boys]

Because it's likely that the 50L/min episode was caused by mouth leakage, perhaps event detection is partially hampered even though 'relatively' efficient therapy is being delivered. I hope others will weigh in with better or more useful information.

I highly doubt that I was mouth breathing for an hour. I suppose anything is possible, but it's unlikely. My hubby knows he is to wake me up if he notices any leaking or mouth breathing.

I am curious, if you toggle between Total Leak and Unintentional leak on your reports, what is the difference?

Here's my data, both ways, for the last two weeks. The last two days have been at 15/12 and the prior days were all at 13/10.
Date ---total leak rate--unintentional leak---difference