OA clusters align perfectly with leak spikes

[Jason S.]

In looking at my data, I noticed leak spikes align perfectly with my OA clusters. What does this mean? Are the OA clusters causing me to move and resulting in a leak spike, or are OA's being recorded due to a pressure loss from a leak?

[Pugsy]

How big is the leak? I had this last night with essentially no leaks and I have had the exact same thing with some medium sized spike leaks but only spikes (no large leaks and not prolonged). I have wondered the same thing myself. With these charts we cannot see exactly which came first. Does Encore Pro let us break down the time line more? I should be getting Encore Pro shortly.

[Velbor]

In looking at my data, I noticed leak spikes align perfectly with my OA clusters. What does this mean? Are the OA clusters causing me to move and resulting in a leak spike, or are OA's being recorded due to a pressure loss from a leak?

With only the information you've provided, there's no way to tell what's happening. Given your past posts speculating about possible "false positive" results, I wonder whether you're thinking along these lines? Another potential explanation is that, in response to OA's (assuming you're in AutoPAP mode), pressure is increasing, and increased leak is often associated with increased pressure.

There also must be the caution that we're built to look for and to see correspondence - even when it may not really be there. The time-scale compression of the Encore reports makes the identification of "alignment" a difficult endeavor at best.

Regards, Velbor

[Muse-Inc]

...

OK, tell me if I've got this right:

• Leak causes a loss of pressure that compromises CPAP therapy that may result in an apneic or hypopneic event.

• A flow limitation or snore or depending on your xPAP the current pressure will cause an increase in pressure that may result in increased leak.

• At this time, users are unable to definitatively associate leak to A or H or increased pressure --> A or H because available software does have sufficient granularity to enable identification of which occurred first.

If true, it seems to me there is a huge gaping hole in optimizing therapy. Leak is a huge issue for us mask wearers and knowing if leaking occurs before or after events seems critical to understanding leak numbers and the overnight effectiveness of therapy.

[Pugsy]

For me I don't think the leaks are causing the events. I have similar leaks and no events also. The leaks aren't huge.
For this one I didn't even have a huge pressure increase to maybe disrupt the seal. One of those things I wonder about in the back of my mind but unless I had glaring leaks and large numbers of events, I don't spend a lot of time on it. I have had reports where I have had some pretty good leaks and not any more events than usual and I seem to "feel" the same. So I think the reporting seems accurate. I can pretty well tell when I get up in the mornings what my report will show by the way I feel.

I would tend to think that in my case that the events make me restless and that may cause a brief increase in the leak.

[Velbor]

OK, tell me if I've got this right:
Leak causes a loss of pressure that compromises CPAP therapy that may result in an apneic or hypopneic event.
....
Leak is a huge issue for us mask wearers and knowing if leaking occurs before or after events seems critical to understanding leak numbers and the overnight effectiveness of therapy.

That's not the way I would express it.

xPAP machinery is designed precisely to AVOID leak resulting in a drop in pressure.

Leak has to be VERY severe and prolonged before it will overwhelm the machine's ability to compensate and maintain pressure.

A leak that severe is likely to awaken the user, with large volumes of air blowing noisily where it's not supposed to go.

Dare I say it .... concern about leak may, in a sense, be grossly over-rated. At the beginning, when fitting a mask and adjusting the straps and learning to sleep with it on, control of leak is crucial. Identifying LARGE mouth leak - which may dry out respiratory passages and thereby increase resistance and cause myriad other problems, even if the machinery can compensate and maintain pressure - and evaluating the effectiveness and the effect of efforts to control it, is crucial.

Leak does need to be watched. But I often see what I think to be obsessive fussing over leak which is often well within machine design parameters and which does not produce clinical symptoms. Worrying about such leak - please, please pardon me - should not keep us awake at night. Velbor

[Muse-Inc]

...Leak has to be VERY severe and prolonged before it will overwhelm the machine's ability to compensate and maintain pressure....At the beginning, when fitting a mask and adjusting the straps and learning to sleep with it on, control of leak is crucial. Identifying LARGE mouth leak - which may dry out respiratory passages and thereby increase resistance and cause myriad other problems, even if the machinery can compensate and maintain pressure - and evaluating the effectiveness and the effect of efforts to control it, is crucial...often see what I think to be obsessive fussing over leak which is often well within machine design parameters and which does not produce clinical symptoms. Worrying about such leak - please, please pardon me - should not keep us awake at night.

Thanks! Now I won't lose sleep over leak

[roster]

......
A leak that severe is likely to awaken the user, with large volumes of air blowing noisily where it's not supposed to go.

Dare I say it .... concern about leak may, in a sense, be grossly over-rated. ......

Finally I can agree with someone on something today. Velbor's comments outline the situation well.

When I look at that leak line in Pugsy's post, I think her leak is under excellent control. But I have often seen lines like that posted and the comments were, "You need to get that leak line as flat as possible."

My personal experience has been that anything at or below 75 lpm (total leak) still gives me a very good therapy. Above 75 lpm I can't remember ever seeing. If it gets that high maybe the mask has fallen off and hopefully I will wake up.

Jason S., We might be able to comment specifically if you would post some of the graphs.

[jnk]

In looking at my data, I noticed leak spikes align perfectly with my OA clusters. What does this mean? Are the OA clusters causing me to move and resulting in a leak spike, or are OA's being recorded due to a pressure loss from a leak?

With only the information you've provided, there's no way to tell what's happening. Given your past posts speculating about possible "false positive" results, I wonder whether you're thinking along these lines? Another potential explanation is that, in response to OA's (assuming you're in AutoPAP mode), pressure is increasing, and increased leak is often associated with increased pressure.

There also must be the caution that we're built to look for and to see correspondence - even when it may not really be there. The time-scale compression of the Encore reports makes the identification of "alignment" a difficult endeavor at best.

Regards, Velbor

I would also think that the first obstruction itself might be enough to spike pressure just enough to start a leak, depending--right, Velbor? Or no?

Just wondering.

jeff

[Jason S.]

OK, here is my graph. Which came first, the chicken or the egg?

[Jason S.]

With only the information you've provided, there's no way to tell what's happening. Given your past posts speculating about possible "false positive" results, I wonder whether you're thinking along these lines? Another potential explanation is that, in response to OA's (assuming you're in AutoPAP mode), pressure is increasing, and increased leak is often associated with increased pressure..

I am using APAP but I'm cant correlate OA's with increases in pressure. My data doesn't support that is happening. I don't think that would account for the spikes.

[Velbor]

OK, here is my graph. Which came first, the chicken or the egg?

I would also think that the first obstruction itself might be enough to spike pressure just enough to start a leak, depending ....

....sufficient granularity to enable identification of which occurred first....

Thank you for posting the graph, Jason. Makes things a lot easier to talk about, even though answers to the questions may still be elusive. The few OA and FL events did not result in notably significant pressure changes which might be expected to result in increased leaking. You may have been right in your first post in speculating that the leak spikes were the result of movement, though I would still wonder whether OA's (1) immediately preceded and (2) therefore necessarily "triggered" the movement.

Let me first address "granularity". The Respironics Encore system presents pressure, leak and respiratory event information for 30-second time windows. (By the way, the ResMed ResScan system - without ResLink - presents data for 60-second time windows.) We have no access to more precise timing than that. Also, I have never seen a satisfactory answer to WHAT information we are being given for that time window. Initial value? Final value? Average or median value? Maximum value? Both the timeframe, and the uncertain content of the data, tend to "blur" things.

Then there are the graphs. The Respironics graphs, as I've commented, are quite tightly time-compressed, so it's quite difficult to accurately draw conclusions about sequence from one graph to another. (By the way, ResMed Rescan very nicely allows for varying the time-scale, as well as the presentation order of the graphs, making such sequencing issues easier to resolve.) Whether Jason's OA events and pressure spikes "align" with each other, to my eye, is inconclusive.

Still, data to help answer the question IS unavailable; but it would require some digging, Jason. In XP, Encore creates files which are left in C > Neo > SCDecodedFiles. These .xml files are used by Encore in creating the report. The files are named decodedSC.xml{mmddhhmm}.xml, where {mmddhhmm} is the date and time the report was run (if two digits are not needed, only one is used). Presumably you ran the report on 11 September, so the file would be decodedSC.xml911hhmm.xml. You can double-click on it to open it as text (your computer may warn you that it is displaying, but not running, the content). Then you need to scroll down, find the relevant times as displayed in the graph, and see what the temporal sequence of events and pressures and leaks actually is. (This may be much more complicated if you ran the report to cover a period of more than one night, or if there was more than one session for the one night.)

Even if the events do "line up", concurrence does not prove causality. You'd need to have (and you should want) a lot more evidence before drawing that conclusion. Instances of OA which do NOT correspond with leak spikes would also need to be explained. All that you might end up with is that "sometimes" instances of OA are "associated with" instances of leak spikes. Unfortunately, that isn't a very satisfying explanation of anything.

Occasional sharp, short "leak spikes" are not uncommon, and very likely are not particularly significant. Regards, Velbor

[Jason S.]

I can't help questioning the validity of the data. Again I had a night where I laid awake for a long time and recorded 5 distinct OAs. Yes, I can throw that data out, but it really calls into question OAs that are recorded when I am asleep, especially when they show a nearly 1.0 correlation with leak spikes. I might even argue the only reliable indicator of a night's therapy is how one feels. Obviously good numbers are good, but are bad numbers always bad?

[Wulfman]

I can't help questioning the validity of the data. Again I had a night where I laid awake for a long time and recorded 5 distinct OAs. Yes, I can throw that data out, but it really calls into question OAs that are recorded when I am asleep, especially when they show a nearly 1.0 correlation with leak spikes. I might even argue the only reliable indicator of a night's therapy is how one feels. Obviously good numbers are good, but are bad numbers always bad?

Depends on how "bad" they are and what may be causing them. People don't breathe the same when they're awake as they do when they're asleep. As I've said many times before, analyzing the data in this therapy is all about "averages".
The "numbers" in that report you posted are pretty good, actually. Trying to determine what caused what is kind of academic (and sometimes frustrating if you let yourself fall into that trap). Did the leaks cause the events or did your movements during the events initiate the leaks?
When using an Auto in a range of pressures, the machine is always going to respond to leaks, snores and flow limitations before it responds to apneas and hypopneas.


Den

[Jason S.]

That 1.8 was a pretty good night for me. Mostly they have been between 2.5 and 3.5 lately. I posted that night because the correlation between leaks and OA's was striking. The night I recorded 5 OA's while awake I didn't have the associated leaks.