Why all the fuss about Leak values?????

[Laurie1041]

Mine are going up. I had one night at 1.3. My average leak is 18L/min. Taping my mouth is not an option. I cannot be sure, but I don't think this is where the problem is coming from.

I feel air escaping from the nasal pillows from time-to-time during the night. When I pull the back of the headgear down, this solves the problem (for the moment). I am using xtra small pillows and have a set of small pillows that I can try. It just seems that the headgear slides or slips on my hair. Any ideas other than mouth taping are highly desired.

I am assuming that I lose therapeutic pressure if I have a leak - like a tire with a slow leak ends up as a flat tire? Some of you have said that on the S9 that it is calibrated for a variance in pressure and should therefore read 0.0L/min. What do my esteemed PAP colleagues recommend? Thanks in advance, Laurie

[jnk]

Mouth leak is bad news, so make sure it isn't that.

http://www.resmed.com/us/clinicians/com ... clinicians

[HoseCrusher]

Laurie, check the hose that goes from the mask to the larger hose. As my use rname implies, I have gone through several of those hoses. It starts with a very small pin hole leak and goes on from there.

A way to check is to plug the end of the hose and turn your machine on and hold the hose up to your face. You can generally pick up even the small leaks this way.

In 6 months I have gone through 4 of those small hoses.

On a side note... my machine reported AHI drops drastically when my hose is damaged and just starting to leak.

[robysue]

Mine are going up. I had one night at 1.3. My average leak is 18L/min. Taping my mouth is not an option. I cannot be sure, but I don't think this is where the problem is coming from.

Is that "average leak" number the one on the S9's LCD in the morning? If so, then it's important to realize that it is NOT the "average" leak value in any ordinary meaning of the word "average".

The Leak number on the S9's LCD is the 95 Percentile Leak Rate Value. The definition of 95 Percentile (95% for short) of a given set of data is the value that is exceeded in only 5% of the cases. So the way to understand the meaning of:

95% Leak Rate = 18L/min

is to understand that for 95% of the time the machine was running, the leak rate was measured AT or BELOW 18L/min. And it also follows that for 5% of the time the machine was running, the leak rate was measured AT or ABOVE 18L/min.

To turn this into something that actually makes sense, you need to realize that 5% of ONE HOUR is equal to 3 minutes. Hence, if you used the machine for 7.5 hours and the LCD screen says your Leak is 18L/min, what that really means is this:

Your leak rate was AT or ABOVE 8 L/min for a total of about

(3 minutes/hour) x (7.5 hours) = 22.5 minutes (or a bit more than a third of an hour)

And hence, your leak rate was AT or BELOW 8 L/min for a total of about

(7.5 total hours) - (0.3 hours with high leak rate) = 7.2 hours with leaks AT or BELOW 8 L/min.


Now, is this kind of leak significant? Well, it really depends much more on the median leak data (which is only shown in ResScan) and the shape of the detailed Leak Rate graph (which is shown only in ResScan).

For example:

Let's assume that CPAPer A and CPAPer B both use their S9 for 8 hours one night. Note that 5% of 8 hours is 24 minutes.

And now suppose CPAPer A is leaking at a rate of 7.5 L/min for most of the night around, and in addition to all the "little" leaks all night long, Patient A also has one 24 minute leak where the leak rate is above 8 L/m the whole time and might even max out at something really horrible----say 55 L/min. In this case, CPAPer A's S9's LCD Leak number will be 8 L/min because for all but 5% of the time, CPAPer A's leak rates were (just) below 8 L/m. But also note that CPAPer A's median leak will be something pretty close to 7.5.

Now suppose CPAPer B has no leaks for most of the night---in other words, for most of the night, CPAPer B has a leak rate equal to or pretty close to 0.0. But CPAPer B also has two 12 minute leaks where the leak rate is just over 8 L/min in both leaks. Since CPAPer B's leak rate is AT or ABOVE 8 L/min for 5% of the time, CPAPer B's S9 LCD Leak number will be 8 L/min. But CPAPer B's median leak rate will be pretty close to 0, and in fact might equal 0.0 if CPAPer B had no detectable leaks for at least four hours out of the 8 the machine was in use.

Now note, CPAPer B probably doesn't have a serious enough leak problem to worry about (unless those two leaks woke CPAPer B up). But what about CPAPer A? Well, it seems to me that if you're leaking all night long at 7.5 L/min, there might just be a real problem. Maybe with mouth breathing? Maybe with the mask not being properly fitted? Pinhole leak in the hose? Maybe a combination of these?

But it's impossible to tell if there is a problem based solely on the 95% Leak Rate that's shown on the S9's LCD screen. You need to see the detailed Leak Rate graph in ResScan to figure out if your leak rate is most usually down close to 0 and there are one or more shortish leaks that are pushing the 95% above 0 or if you're leaking all night at a rate that's high enough to affect how you feel the next morning.

I feel air escaping from the nasal pillows from time-to-time during the night. When I pull the back of the headgear down, this solves the problem (for the moment). I am using xtra small pillows and have a set of small pillows that I can try. It just seems that the headgear slides or slips on my hair. Any ideas other than mouth taping are highly desired.

Let me guess, you're using a Swift FX for Her aren't you? I'm currently using a Swift FX for Her myself. I'm swapping back and forth between XS and S pillows, but I'm not sure that matters that much for me. However if the XS pillows are small enough to be going to far into your nostrils, the trying the S pillows will probably help with getting a better seal.

As for the backstrap of the headgear sliding around: The most common solution to that problem seems to be to sew a bit of velcro (the stiff side with HOOKs) onto the part of the backstrap that lies against your hair. The velcro should gently catch your hair and stabilize the position of the backstrap and help maintain the seal. I have not yet done this with my Swift FX backstrap----it's one of my projects for this weekend though.

I am assuming that I lose therapeutic pressure if I have a leak - like a tire with a slow leak ends up as a flat tire? Some of you have said that on the S9 that it is calibrated for a variance in pressure and should therefore read 0.0L/min. What do my esteemed PAP colleagues recommend? Thanks in advance, Laurie

As I said above, you can't tell from the 95% leak rate whether you've got a long, slow, serious leak or whether you had a shortlived leak that didn't really affect the vast majority of your therapy the night before.

And yes, the S9's software is set up to automatically subtract the expected leak rate for the exhaust flow from the raw leak rate data so that the user doesn't have to do the subtraction his/herself. Of course that does mean you need to make sure that the mask setting is correctly set for your type of mask. And while it's nice to aim for a 95% Leak Rate of 0.0, I doubt that many folks actualy get a 95% Leak Rate of 0.0 night after night for weeks and months on end. For my own data----I know I can reliably expect my median leak rate to be 0.0 every single night. But my 95% leak rates vary from 0.0 to about 12L/min. Most nights my 95% leak rate is somewhere between 3 L/min and 10 L/min and I've decided not to worry about it at all because the detailed graphs show that my leak rate is usually a flat line at 0.0 except for a few (1 to 5) short leaks on the nights. And the total time spent leaking is (well) less than a half an hour each night.

[nanwilson]

To keep the headgear from sliding on you hair just add two bits of velcro (the scratchie bit not the fluffy side) to the inside of the bottom back strap. The velcro kinda stops the straps from sliding but doesn't stick in your hair. Make sure its the scratchie side of the velcro not the smooth side.
Nan

[Laurie1041]

Mouth leak is bad news, so make sure it isn't that.

http://www.resmed.com/us/clinicians/com ... clinicians

Hi jnk,

Thank you for the link. I don't know how to detect if I have a mouth leak. My mouth is not dry in the mornings. The only witnesses I have are my two cats and trying to get anything out of them is next to impossible.

I'll try the recommendations using Velcro to keep my headgear in place and switch out the pillows and see if I can maintain a better seal. I'll know if my seal is better by not being woken up with air shooting out of my nostrils.

I'll cross the path of a FFM if my downloaded data is clearly showing significant leakage that will interfere with my therapy. Thank you again for the information. Hugs, Laurie

[Laurie1041]

Laurie, check the hose that goes from the mask to the larger hose. As my use rname implies, I have gone through several of those hoses. It starts with a very small pin hole leak and goes on from there.

A way to check is to plug the end of the hose and turn your machine on and hold the hose up to your face. You can generally pick up even the small leaks this way.

In 6 months I have gone through 4 of those small hoses.

On a side note... my machine reported AHI drops drastically when my hose is damaged and just starting to leak.

Hi HoseCrusher,

Thank you for responding. I have read on this forum that there are problems with the S9 Slimline and Climateline tubing. I will definitely check the integrity of my hose. Great idea. Thanks again! Laurie

[M.D.Hosehead]

The CPAP Wiki article on leaks is very informative.

wiki/index.php/Mask_Leaks

[Laurie1041]

[/quote]Is that "average leak" number the one on the S9's LCD in the morning? If so, then it's important to realize that it is NOT the "average" leak value in any ordinary meaning of the word "average".

The Leak number on the S9's LCD is the 95 Percentile Leak Rate Value. The definition of 95 Percentile (95% for short) of a given set of data is the value that is exceeded in only 5% of the cases. So the way to understand the meaning of:

95% Leak Rate = 18L/min

is to understand that for 95% of the time the machine was running, the leak rate was measured AT or BELOW 18L/min. And it also follows that for 5% of the time the machine was running, the leak rate was measured AT or ABOVE 18L/min.[quote]

Hi,

Thanks for your thoughtful response. I understand the mathematical concept. I have downloaded and installed ResScan 3.11 and will upload my smart card and begin looking at trends. First, I need to fix my nasal pillow/headgear dilemma. Thanks again! Laurie

[jnk]

. . . I don't know how to detect if I have a mouth leak. My mouth is not dry in the mornings. The only witnesses I have are my two cats and trying to get anything out of them is next to impossible. . . .

Cats are like that. They know, but they ain't tellin'.

For serious big-time trouble-shooting of leaks, nothing beats the software for seeing charts of the leak.

[Shellie_p]

On mine, I added an extra strap (got the idea from this site under the labrat awards) that actually goes around the back of my neck and went to the bigger pillows then the RT recomended. (was recomended to me to use the small but med works best) And haven't had any leaking issues at all from Swift. The back of the head strap never stays in place for me. So a couple moments with 2- 1inch sets of velcro and a few inches of Black 1 inch elastic was all the extra I needed.

[cwied]

Mouth leak is bad news, so make sure it isn't that.

http://www.resmed.com/us/clinicians/com ... clinicians

Hi jnk,

Thank you for the link. I don't know how to detect if I have a mouth leak. My mouth is not dry in the mornings.

When I have mouth leaks, I wake up with my mouth open and dry and my nose feels dry and painful on the inside. It's possible that I'm just particularly sensitive to a dry nose, but I would think you would feel it if you have the problem.

The interesting thing about the link that jnk provided is that it explains the mechanism of losing therapy:

Increased nasal symptoms can worsen OSA and lead to increased nasal airway resistance.

The corollary to this is that if you aren't getting a stuffy nose from your leaks, you aren't worsening your OSA. Given that the machine will compensate for leaks by increasing the airflow to maintain pressure, the leaks may not matter. I would say you should only worry if your AHI numbers are going up with the leaks.

I am assuming that I lose therapeutic pressure if I have a leak - like a tire with a slow leak ends up as a flat tire?

No - the blower increases the flow volume to maintain the pressure at its output hose at the therapeutic level. With the higher flow, the pressure difference between the blower and the mask will increase, though. I assume this is where the 24lpm red line comes from. If you're below this level, the difference is presumably not therapeutically significant.

I have a Swift FX and am fighting leaks as well. On the other hand my AHI has been consistently below 1. I've just started using a chinstrap, and it looks like it will help. Last night my 95th percentile leak was only 8.6 (it's been as high as 24 without the chinstrap). My nose feels a lot better with the lower leak numbers, so I think I won't worry about the remaining leaks.

[GumbyCT]

Is that "average leak" number the one on the S9's LCD in the morning? If so, then it's important to realize that it is NOT the "average" leak value in any ordinary meaning of the word "average".

The Leak number on the S9's LCD is the 95 Percentile Leak Rate Value. The definition of 95 Percentile (95% for short) of a given set of data is the value that is exceeded in only 5% of the cases. So the way to understand the meaning of:

95% Leak Rate = 18L/min

is to understand that for 95% of the time the machine was running, the leak rate was measured AT or BELOW 18L/min. And it also follows that for 5% of the time the machine was running, the leak rate was measured AT or ABOVE 18L/min.

Hi,

Thanks for your thoughtful response. I understand the mathematical concept. I have downloaded and installed ResScan 3.11 and will upload my smart card and begin looking at trends. First, I need to fix my nasal pillow/headgear dilemma. Thanks again! Laurie

You had the "/" in the wrong spot should be before the 2nd quote. You can edit your post if you want.
HTH

[jnk]

. . . The interesting thing about the link that jnk provided is that it explains the mechanism of losing therapy:

Increased nasal symptoms can worsen OSA and lead to increased nasal airway resistance.

The corollary to this is that if you aren't getting a stuffy nose from your leaks, you aren't worsening your OSA. Given that the machine will compensate for leaks by increasing the airflow to maintain pressure, the leaks may not matter. I would say you should only worry if your AHI numbers are going up with the leaks.

My understanding is that the way you lose therapy with mouth leaks is that the therapy pressure escapes out of your mouth so that the OSA worsens from the lack of therapy pressure. Machines can compensate for leak up to a point, but mouth leak is often too severe for a machine to compensate for.

The link I gave is a little confusing in how it mixes general points about "mouth breathing" with its points on the more serious problem of "mouth leak."

One problem with attempting to use AHI to judge the severity of leak is that machines cannot accurately sense your breathing events if the leak is too large. Mouth leak can cause events that never get recorded during the leak.

The other problem with leak is that the noise and discomfort of leaks can fragment some people's sleep, completely independent of breathing-event considerations. Leaks can tickle, squeak, and pop, depending on the mask, and those intermittent changes in sounds can cause arousals.

I agree that some make too big a deal about leak. But it is just as possible to make too little a deal about them, too. IMO.

[cwied]

My understanding is that the way you lose therapy with mouth leaks is that the therapy pressure escapes out of your mouth so that the OSA worsens from the lack of therapy pressure. Machines can compensate for leak up to a point, but mouth leak is often too severe for a machine to compensate for.

I definitely could see that that would be an issue if the leak is big enough. However, given a 95th percentile leak value under 24, I had assumed that the leak was within the ability of the machine to compensate.

One problem with attempting to use AHI to judge the severity of leak is that machines cannot accurately sense your breathing events if the leak is too large. Mouth leak can cause events that never get recorded during the leak.

I've seen this statement a couple of times before, but I'm not sure I quite understand why this would be. Is it that the amplitude gets harder to judge when the leak is high? I've looked at the waveform when I have my mouth open, and I'm pretty sure that the event detection is accurate in my case. Then again, the only events the machine typically detects for me are central apneas, so I don't know if it would be different for people who have hypopneas while on CPAP.

Here's an example of event detection while my mouth is open. It's also interesting to see how the waveform changes when I close my mouth (waveform is skewed to the left due to the ResScan bug):