My PSG sleep study - 2 years after starting CPAP treatment

[rested gal]

Thanks to sleepydave (manager of a top sleep lab), I finally had a full PSG study! Three nights of being wired up, October 17 - 19, 2005. What a cool learning experience! My account of it is here:

First Sleep Study...finally! - rested gal

Full reports from all three nights of PSG

[dizzy]

Looks like you had a pretty good crew, my first study I was worried they were going to steal the wheels of my car in the parking lot

I read your 3 studies, interesting #2 showed in non-supine positions you had 0.0 events (seems that is where you should try and sleep).

On your notes with the 420e, I agree without the exact delivery hose that came with the machine and internal 1/8" pressure sensor line I would think your results from it would be so skewed that the results wouldn't be of any use. Even a small change to the length of the stock delivery hose would impact the machine. Then some hoses are smoother on the inside and even this can impact delivery as it wasn't calculated in on the machine's calibration. In other words it is expecting to see xx volume flowing through the hose as you breathe and static feedback pressure from that pressure sensor hose.

When you add any length to the hose you add static resistance into the circuit. On that machine it uses that pressure sensor hose (the small clear one inside the delivery hose) to measure the pressure right at the mask. Most other machines have hence done away with that hose as you can calculate any resistance if the delivery hose is known and come to the same conclusion with less hazard to the machine. The hazard is that little hose being a water-main directly back inside the machine, any water in the mask for whatever the reason say from rainout that gets in that hose can go inside the machine under whatever cpap pressure your at. Without that hose and on other machines any moisture is always blown away from the machine. What you notice it one day when it don't turn on

Back to looking at your Study2titration, boy it's a wash between 8cm or 9cm being your ideal pressure. BUT, while it shows you still had 7 hypos at 9cm pressure none were obstructive by nature, but at 8cm you slept for 77minutes and 24 of that was in REM (which is much better in my opinion, more dreaming), yet at the higher 9cm pressure, you slept a much longer period but had even less REM sleep for 3x the same sleep period, in fact only 3 minutes worth. While your SAO2 levels were a tad better at 9cm, that could have been from shallower breathing due to more REM you had at the prior pressure. Seems at 8cm pressure your SAO2 levels never get down to significant levels to worry about.

In other words that extra 1cm pressure may be what was keeping you from going back into REM. Then as the study progressed most people seem to experience more REM during the early morning hours, time may be shorter with each you have but most have more. I'll stick with my guess that the 8cm is your ideal pressure.

What did they say about those Spontaneous Arousals seen? If they were like mine, they kinda mumble under their breath (um...we really don't know what the cause of those are... ). Hint: your report indicates loud snoring (join the club) these can interrupt your sleep and be the source of those arousals.

I don't know much about those PLM or leg movements, I know they can be related to arousals but is that always bad? I didn't correlate these to any titration pressure. I just looked at what I seen that delivered the most deep & REM sleep, because those seem to be what makes you feel refreshed in the morning.

By comparison, how did you sleep/feel with the Remstar Auto? My experience with it (looking at your 2nd study) is it would have backed off the 9cm pressure because the shear # events seen at 9cm were higher.

thanks for sharing.

[deltadave]

Actually, I didn't think the resistance of the tubing would make a big difference in the measurements, resistance is more related to flow, and there really isn't too much flow in the pressure line, it's more like transmission of pressure, which is fairly instantaneous.
I thought the readings might be dampened by the bulges in the argyle tubing.
However, to check this after the fact, I took a 6-inch length of 420e style tubing, and 8 feet of argyle tubing, including twice as many bulges as the monitor length, and ran a simultaneous test of the waveform characteristics. Tried to get in a little FL style breathing, too. This is what they look like:



Right, does it make a difference if I tell you which is which?

Argyle tubing is quite rigid, and designed specifically to resist the tendency to kink. Not that it can't, and without reading the 420e transducer directly we don't know exactly why it gave the readings it did. Perhaps we can look at the areas that the 420 were calling CAs to see if there are some clues.
The key is not why it saw apneas, but why it thought it saw cardiac oscillations.
deltadave

[deltadave]

Oh, right, the spontaneous arousals. There really shouldn't be a truly "spontaneous" arousal, you kinda hafta to figure out why.

Here's some of the spontaneous arousals (yes, this is the same set of epochs from the CSDB thread) but the arousals are noted with the red arrows. You can see in the breathing or flow channel how the tops of the waveform flatten out immediately prior to at least two of the arousals. These are the infamous flow limitations.



The reduction in flow (actually, it's really pressure, but that's being picky) is not enough to call it a hypopnea (it doesn't drop to less that 50% of baseline with this arousal, or, using the Medicare standard, there's no desaturation) but if you had enough of these, you would call this a case of Upper Airway Resistance Syndrome (UARS).
deltadave

[rested gal]

dizzy, initially I thought the same thing you did....that using the Argyle tube instead of the standard tube for the 420E was what made my autopap perform so atypically (for me) that night in the sleep lab.

When I got home, I removed the Argyle tube, intending to throw it away. Then I decided, "What the heck....let me try the Argyle one more time to collect one more night of crazy data." I threaded the Argyle back down the main air hose again.

Well, wonder of wonders, my data the next morning was normal again. Exactly the same kind of results I had been getting with the 420E when using the official sensor tube. I've left the Argyle line in place ever since. It's been a month now and the machine has behaved without a hitch. Still using it with the Argyle tube. No problem.

By the way, the reason I even used a makeshift sensor line at all at the sleep lab and here at home, was so I could use my heated hose. The heated hose I have is one I was using with another machine (Respironics) thus did not have a sensor line included.

[dizzy]

The 420e is equipped with a more modern pneumotach sensor and theoretically can detect cardiac oscillations. However, I have my doubts if it can do so accurately. The Remstar Auto and the Devilbiss 9000 also have a similar pneumotach sensors. The Resmed Spirit uses a more primitive hall-door effect type sensor and relies on its A10 algorithm to limit pressure increases to 10cm. There was/were more details on this in TAS archives in past discussions with Perry, SWS and others going back several years now.

Which graph has the more resistive tube in circuit?

I'd say if there is a difference it would be the bottom graph or cpap2, it appears to have more flattening compared to the top graph. While it doesn't appear to change the sensitivity of your instruments to read much it could easily do so on the 420e's ability to read and respond to events especially if that pressure sensor tube is missing or not in the circuit. Volume vs pressure. These machines work off of pressure. In order to know the volume or flow you have to know the static pressure and use a precisely calibrated pitot tube then calculate out the volume. Some in-line pneumotach sensors such as those from Medical Graphics used in exercise training masks have these pitot tubes built-in, cpaps do not because volume is not so critical to their operation. The volume is set within. There is only so much volume that can flow through that diameter hose at a given pressure anyway.

When you extend the length of the hose, change diameters and/o induce bends and/or use a more resistive inside surface hose in the circuit it can add static resistance and to the ability of the machine to read and respond to events. This is why some manufactures only certify their machines for use with their own interfaces. Why did the 420 see events the other equipment didn't? You really don't want me to answer that

RG: you indicated the 420e picked up centrals where the monitoring equipment didn't pick these up? Theoretically what happens when a CA is seen? isn't it where the patient appears to stop breathing. If the 420 had difficulty reading your breathing pattern due to the longer delivery hose I would think it would have been seen as a CA.

[deltadave]

Actually, the NPSG picked up all the events and then some. It picks everything up, and then the scorer must decide what is significant and what is not. Guess the 420 does the same thing, now that I think about it.

Based on the download from the 420e, here's where the AutoPAP was ticking off events. There appeared to be a group of centrals, if I am reading this thing correctly, at about 2.25-2.75 hours into the study:



Looking at the NPSG, I think these are some of the events the 420e was calling CAs, or at least hypopneas:



A hypopnea has 50-80% reduction from baseline, apneas beyond that, so a couple of these things I wouldn't have a problem calling apneas. They still had some flow limitation quality, so we stuck with hypopneas. For that reason, and also because they appeared to have some almost central quality to them. A reduction in response to some brief hyperventilation. And I say this because this is what was happening immediately prior to this:



PLMs. Leg kicks that generated arousals, brief periods of hyperventilation followed by normal breathing. And this is what I really thought would mess up the 420 algorithm. Didn't as much as I thought it would. Perhaps with IFL1 off, it didn't react as severely as it could have. The pressure waves were all pretty much flow-limited, and it does look like the 420 did respond somewhat, figuring some of these were hypopneas.
Now, since the cluster of hypopneas occurred right after the group of PLMs, you gotta kinda wonder if they were simply a response to more PLMs. The frequency of the PLMs has dropped off somewhat, but I think you can make an easy case for a few more PLMs, with the resultant breaths just happening to be flow-limited.

We can talk about this now, give thought to what has happened, and reconsider therapies. But that's the occasional problem you run into with an AutoPAP and a tick on a graph. What's that tick really mean?

So the problem is, without directly tying into the 420 transducer, there's no real good way to know what it was actually seeing. I used PTAF2 transducers, which are in widespread use, pretty much considered to be a good transducer, and I felt our data to be quite good. I had considered tying directly into the 420, but if a circuit board got smoked there definitely would have been a HostileGal around.

With no true apnea, tho, why the 420 was thinking it saw cardiac oscillations is beyond me.
Maybe next time...
deltadave

[rested gal]

[dizzy]

RG wrote:

[deltadave]

Even PB refers to the Rappaport study re: accuracy of the cardiac oscillation function as only being about 60%:

Cardiac Oscillations on Airflow Signal in CPAP

So don't say "Omigod!" when you see 3 central apneas. Maybe if you see 100, then you can refer to the thread on Complex Sleep Disordered Breathing (seriously)(sort of).

Yeah, I really don't know the technology behind the 420 pneumotach. If it's simply measuring a pressure gradient (and not really flow, as in some of the other technologies) maybe that's why we can get away with fiddling with the pressure tube characteristics.

Changing the length of the tubing changes resistance linearly, but changing radius of tubing changes resistance to the power of 4. Which kinda makes me wonder about that little tube in the first place. Seems like the signal to noise would be horrific compared to just measuring it from the patient tube. Yet if the signal is is simply amplified to generate a "flow" waveform, and our little experiment with the argyle tubing shows that you might get away with that, maybe it's academic.

If that little tube is only measuring part of the gradient, or pressure, and there is in fact no flow going through there, that might explain things.

If you're hooked up to PSG with 2 external transducers like we were, maybe you can do this stuff. If not, I wouldn't experiment with tubing lengths and diameters.

BTW, if changing the radius of a system changes the resistance exponentially, keep that in mind when you change mask interfaces, especially to those nasal pillow systems.
deltadave

PS - You really see a difference in those 2 waveforms?

[rested gal]

LOL, dizzy, I love ya to pieces, but you do tend to let your poor experience with the earliest PB 420E's color how you've regarded that good machine ever since. I don't doubt that the first ones off the assembly line had problems. I'm glad quality control kicked in before I got mine.

Doorstop? heheh, cute notion, dizzy, but I think of it as a Brick! A good solid brick giving me good treatment every night. Oh, I've used several concrete block size autopaps too. Got the same good treatment from any of 'em.

I don't know how many ways I can say it - my 420E's pressure swinging behavior on that third night at the sleep lab was NOT TYPICAL of the way it has been treating me before that night and has been treating me after that night. That night was an odd performance, for whatever the reason.

The only quibble I'd ever have with the 420E is that its IFL1 trigger (on by default) needs to be turned off for some people. Most probably are fine with both IFL1 and IFL2 on. Thanks to -SWS looking at my data, I found out IFL1 definitely needs to be "off" for me. I can think of at least five other pretty savvy message board posters who also reported having to turn IFL1 off on their machines. It can be a troublesome trigger for some people, if left on.

I've posted quite a few of my SL3 graphs showing not only the sleep lab night, but also how the 420E normally treats me (with IFL1 off) as well as couple of graphs showing how out of control it behaves if I have IFL1 on.

I've added some info on the pictures making it easier to understand what the settings were in each graph:
tnlc.com/Lara/laura/osa/study-Oct2005/420E-Screenshots-SL3-Data/

One odd night (at the sleep lab) for a reason that may never be clear (I will always suspect a pinched sensor line on that night) does not, in any way, negate the usual very good treatment I get from the 420E, night after night after night....as long as I keep IFL1 turned off.

And yeah, I'm using it (AND every other autopap I've used - REMstar and Spirit) with the low pressure set up at or very near what would probably be my "prescribed pressure" if I were using a straight cpap machine. All of them work best for me like that, whether I use a brick or a concrete block size machine.

As -SWS remarked once:

"My opinion is not that the 420e is an excellent AutoPAP because it has IFL1, IFL2, and "command on apnea" adjustable parameters. Rather, my opinion is that the 420e is a superb AutoPAP despite having these features that can get so many patients and therapists into trouble."

That was from one of his posts on page 3 of this very interesting old topic:
Jan 11, 2005 subject: Bman: Spirit Overnight Indices

Another interesting long discussion was here, when -SWS and John discussed the 420E's IFL1 and IFL2 triggers:
Apr 10 2005 subject: A question (or two if I can remember) about the PBG 420E

P.S. I didn't see any difference in those two waveforms, Dave.
dizzy, you must have one heck of a magnifying glass.

[rested gal]