Reporting Leaks: My Air v.s. Sleepyhead

Thanks for the information, Pugsy. This is exactly what I was looking for, and I think you nailed it with this:

For this reason I don't get all upset until I see leaks over 35 L/min and for a prolonged period of time. I don't sweat any large leak that just lasts 5 or 10 minutes as long as it doesn't wake me up. So learn to eyeball the leak line graph and how much time is spent deep in large leak territory.
Things don't go totally in the toilet at 25 L/min..it's a gradual worsening of the situation and the further you go the worse it gets.
If 95% of the time you were at or below 24 L/min...I wouldn't give a second thought to what that other 5% went to as long as I slept through it.

I just went back and looked at a sequence my most recent 95% numbers and they looked like this: 29,26,23,29,31,20,23,26,26,20,20. The graphs did show occasional spikes over 35-40, but they were for very short periods of time. Probably when my mouth blew open and I quickly corrected without waking up.

on CPAP (as in one constant set pressure) leaks are not much of a concern.

on APAP, where the machine needs a certain accuracy to determine its own reaction, that is a total different scenario. If you don't care for the auto-reaction and just want some pressure you can ignore pretty much every leak. Big, flat leaks are usually a good indication for mouth-leaks. Some short leaks are most likely changes in sleep-position.

as long as you have leaks you can also ignore nearly every event flagged by the machine (including the flowlimitation graph). Those are just rough estimates (during leaks). In most cases they are more wrong than right.
One approach to such events is "increase the pressure" - helps of course, as you start to actually get some pressure in the range of therapeutic pressure into your airways. You can achieve the same (or even better) with 'just' fixing the leaks.

But as pugsy said: don't worry too much about that. If it doesn't wake you up and you feel rested in the morning everything went fine. (you might however want to fine-tune your lower pressure range if you have near constant leaks - reduces greatly the time you get less-than-actually-needed pressure)
There is also a point where leak-mangement causes more discomfort than it does good.

I don't have time at the moment but I will try to dig out an old report that showed elevated 95% numbers that didn't really mean all that much once I looked at the actual leak graph.
I will have to redo an old image that is held in Photobucket and I just don't have time right now.

I don't really even bother looking at the graphs until the 95% number is over 30 as long as I have slept well. If over 30 I will then look and evaluate.


Image removed.

I am in a bad mood today. Not enough sleep last night. Tired of bullshit.

With respect - but all I(!) see there is: during leaks the machine went berserk. That is actually a very good example of the problem with leaks.

The last "large leak" shows it very clear: the device can no longer measure the flow correctly (that is debateable - most likely that was a mouth leak and thus the flow might be not that much off^^) - it reacts in this case by increasing the PS to max.
The same can be seen on the other leaks as well. In the time before the last leak the PS-increase correlates nicely with the leaks. (as said: increasing the pressure always works - with no leaks I doubt the machine would have increased the PS)

during the time with the 2 large-leaks flags it even looks like the machine totally miscalculated the leak and thus the flow. I would not be surprised, if the baseline of the flowrate is under (or over?) the 0-line during that time. Most be funny to exhale more than one inhales or the over way around.

As I said: not that much of importance if one is well dialed in - total different scenario with a "wide open" APAP-device.

Guest wrote:With respect - but all I(!) see there is: during leaks the machine went berserk. That is actually a very good example of the problem with leaks.

The last "large leak" shows it very clear: the device can no longer measure the flow correctly (that is debateable - most likely that was a mouth leak and thus the flow might be not that much off^^) - it reacts in this case by increasing the PS to max.
The same can be seen on the other leaks as well. In the time before the last leak the PS-increase correlates nicely with the leaks. (as said: increasing the pressure always works - with no leaks I doubt the machine would have increased the PS)

during the time with the 2 large-leaks flags it even looks like the machine totally miscalculated the leak and thus the flow. I would not be surprised, if the baseline of the flowrate is under (or over?) the 0-line during that time. Most be funny to exhale more than one inhales or the over way around.

As I said: not that much of importance if one is well dialed in - total different scenario with a "wide open" APAP-device.

Wasn't debating functionality of whatever in the face of large leaks.
Example was shown just to show 95% numbers that were elevated but the actual time in large leak wasn't all that bad.
That's all it was. 31 minutes total (multiple segments and not just 1 prolonged segment) out of 7 1/2 hours of sleep. Far from the end of the world. I slept fine. I won't lose any sleep worrying about the numbers or whatever. This was a rare "bad" night for me anyway.

I hear and feel you - regarding leaks there is no such thing as the one single number or value.
I started the "whole discussion" just in response to this:

Pugsy wrote:Here's what I happened to have happen to me all in one night so that I could see what the machine did when presented with large leaks that got progressively worse...and what I pretty much live by in terms of what I look at for my ResMed leak.

Leaks over 24 and to 30 L/min...will get a large leak flag but I found that the machine seemed to respond and flag events normally.
Leaks between 30 and 35...I started seeing some unknown apnea flags. The machine knew something was going on but couldn't make up its mind as what to call it.
Leaks over 35 L/min...the machine quit flagging anything when 5 minutes before I was seeing unknown flags. Now I don't for one minute think that nothing happened when 5 minutes before a lot was happening. The leaks got so bad the machine didn't know anything was going on. It lost ability to sense, record and most likely respond when leaks went over 35 L/min.

For this reason I don't get all upset until I see leaks over 35 L/min and for a prolonged period of time. I don't sweat any large leak that just lasts 5 or 10 minutes as long as it doesn't wake me up.

In my opinion that is a bit very "lax".
Taken the night you provided - there was all in all nearly a complete hour when you were "flying blind".
If it would happen to me, I wouldn't loose any sleep over just one night as well. If I would see such "large leaks" constantly, I would do something about that - or if I could not, I would switch to CPAP.

And that was, what I was trying to say all along: leaks must be seen from 2 sides:
1. it's annoying to the patient - or not. Fix it or not.
2. it messes with the ability to measure anything - thus with the ability for any kind of "auto"-CPAP device to react to anything.
It does not really matter how big the leak is - event detection simply is no longer reliable - meaning: all the shiny values like AHI or whatever are simply crap if that happens. The bigger the leak the less reliable everything gets. Clear and reliable distinction between an OA and a CA stops long before 24 liters per minute - the UA flag is just ResMeds way of telling the patient: "holy crap, this so bad I can't hide that from you." (It is not like ResMed is flagging pro active or beeing very talkative at all! - in the end all the care for are flowlimitations - your guess at what leakrate that "detection" starts to get unreliable - you can either count that flow variations off as leak or put it towards the respiratory flow - it is not like real leaks are that constant over time - look at the funny face-farts^^)
If it makes up a significant part of the night (I call 1 hour out of 7 significant) all went sideways.

Personally I make it like Goofproof: my red-line is set for sth. around 10 to 15 liters per minute. (for philips you have to add the intentional leakage - ending up around his numbers - depending on the current pressure)

Guest wrote:In my opinion that is a bit very "lax".

Well, I tell you what. It's not news here that I am a bit lax by some people's standards. And you know what else? I really wasn't asking for your opinion was I?

Guest wrote:Taken the night you provided - there was all in all nearly a complete hour when you were "flying blind".

Nope, not nearly an hour. 31 minutes that you assume I was flying blind. I disagree.

Guest wrote:If it would happen to me, I wouldn't loose any sleep over just one night as well. If I would see such "large leaks" constantly, I would do something about that - or if I could not, I would switch to CPAP.

Didn't happen to you. Happened to me and I already said that this was not something I saw often. And I am not about to change my therapy...at this point it is working quite well thank you very much. If I should ever need your advice I will be sure to not ask.

The image posted was NOT an image to show what I found when I mentioned what I found with the progressive leaks. That's another image held in blackmail by Photobucket and I don't have time to go dig it out.

The image I shared was JUST to show how 95% numbers don't always reflect how the leaks went for the night....that sometimes they make things out to be worse than they really are. It wasn't to ask for opinions about my leak or therapy or whatever. I think I can pretty much get things figured out quite nicely on my own.
That's the ONLY reason I shared it. No other reason...is that so hard to understand?

Thank you so much for your unsolicited critique of my therapy. I will be sure to not remember it.

I apologize for that. I was not trying to critizise you or your therapy at all.
Somehow I was under the impression that this thread was about Comfortably Numb's question about how to interprete leaks.

I apologize again if this is the wrong place to actually discuss such things controversial.

FWIW, just stick with pugsy's approach - that has proven to be working.
Of course YMMV.