How does ResMed S9 machine differentiate OA and CA?

[jcarerra]

Does anybody know how the ResMed S9 machines decide an apnea is a cenral rather than an obstructive?

Obviously it has no input to"see" if chest is "trying" to breathe--so how does it decide that a central is happening?

[zonk]

The link to "Central Sleep Apnea Detection and the Enhanced AutoSet Algorithm" white paper does not work anymore

You can find some info in the clinical manual and there is a nice diagram illustration

From http://www.s9morecomfort.com/s9morecomfo...toset.html
CSA detection: It uses the Forced Oscillation Technique (FOT) to determine whether the airway is open or closed during an apnoea. Small oscillations in pressure are added to the current device pressure. The CSA algorithm uses the resulting flow and pressures to measure whether the airway is open or closed.

[jcarerra]

Thanks.
Do you have any knowledge as to how accurate/successful this discrimination is?

[Robert6]

Adding to Zonk's message (see the 2nd video in Zonk's link),

When there is an obstructive sleep apnea (OSA) event, the measured pressure oscillations are undiminished but the flow oscillations are gone because the airway is closed.

When there is a central sleep apnea (CSA) event, the measured pressure oscillations are diminished but the flow oscillations are undiminished because the airway is open.

Regards, Bob

P.S. I don't seem to have that feature on my S9 VPAP adapt in ASV mode. Possibly because the algorithm may not need to differentiate between OSA and CSA in ASV mode where the EPAP is fixed and the IPAP varies. Maybe if I tried the ASV Auto mode on my machine, it might be there because the machine may then need the FOT to find the EPAP automatically.

[zonk]

ResMed and Philips Respironics engineers are adamant the detection is fairly accurate but the machine does not know whether you,re asleep or awake. From http://www.healthcare.philips.com/main/h...apneas.wpd
By analyzing the flow response to device initiated pressure pulses during an apnea, the Philips Respironics proprietary algorithm is able to reliably distinguish an obstructed airway apnea from a clear airway apnea.

[vsheline]

Maybe if I tried the ASV Auto mode on my machine, it might be there because the machine may then need the FOT to find the EPAP automatically.

Hi Bob, welcome to the forum!

Yes, that would be an interesting experiment. Perhaps you could change to ASVauto mode with Min EPAP of 9.0 and Max EPAP only slightly higher, like perhaps 9.2 or so.

But you might want to wait a couple months if you need to wait until you are past any "compliance" requirements and the insurance company has paid for the machine.

By the way, your Min PS setting is maxed out as high as your machine allows (6). That may be a little too high. I suggest wearing a recording Pulse Oximeter ocassionally while sleeping, to verify your average saturation percentage of oxygen (SpO2) is not higher than around 96% or so. 94-96 is considered ideal, and 90-93 is probably fine, too. Less than 89 is considered low. Less than 75% is considered severely low. An average of 97% or higher (not counting the occasional short dips caused by apneas), if lasting for hours on end, may be unhealthful and can lead to dangerous conditions.

[vsheline]

The link to "Central Sleep Apnea Detection and the Enhanced AutoSet Algorithm" white paper does not work anymore

New link to "Central Sleep Apnea Detection and the Enhanced AutoSet Algorithm" white paper:

http://www.resmed.com/int/assets/documen...-paper.pdf

[archangle]

The first problem is that an "apnea" isn't an apnea if you're awake, and the machine doesn't know whether you're awake. That applies to both central and obstructive apnea. That's why an in-lab PSG sleep test needs EEG leads on your head.

The second problem is that the machine can only "see" whether your airway is clear, it can't see whether your chest is making the effort to breathe. Your airway can collapse during a while your brain isn't trying to breathe. In this case, the machine won't see a "clear airway," so it won't flag it as a "clear airway" apnea. It WILL still know you had an apnea and will probably say "obstructive."

The third problem is that it's sort of a delicate measurement to make. Things like mask type, stuffy nose, leaks, narrow airway, etc. might make a difference. I suspect the manufacturers take the safe approach and don't call it central if they aren't sure.

My conclusions:

You have to live with the "not asleep" problem, it's important, but not relevant to central vs. obstructive. Let's ignore that for the moment

After that, my opinion is that if it says "clear airway," it probably is a true central. The machine will still record an apnea, but it just won't say "central" or "clear airway."

I don't think this is too serious of a problem. It's useful when the machine can tell you it's a central apnea, even if it doesn't flag every central. Just be sure you understand the limitations.

I don't seem to have that feature on my S9 VPAP adapt in ASV mode. Possibly because the algorithm may not need to differentiate between OSA and CSA in ASV mode where the EPAP is fixed and the IPAP varies. Maybe if I tried the ASV Auto mode on my machine, it might be there because the machine may then need the FOT to find the EPAP automatically.

I've heard that the S9 VPAP Adapt will not ever report a central apnea when it's in ASV mode.

[zonk]

The link to "Central Sleep Apnea Detection and the Enhanced AutoSet Algorithm" white paper does not work anymore

New link to "Central Sleep Apnea Detection and the Enhanced AutoSet Algorithm" white paper:

http://www.resmed.com/int/assets/documen...-paper.pdf

Thanks, looking-up and finding it ... good detective work

The enhanced autoset algorithm only available on autoset mode (not CPAP mode or on S9 Elite, set pressure is fixed)

On autoset mode, the machine not only distinguish the apnea whether obstructive or central but also respond differently, no pressure increase if central apnea detected. The autoset respond to snoring and flow limitation, snoring and flow limitation precede an obstructive apnea so the pressure increase is in response to signs of 'closed airways", not "open airways". I,m biased towards the autoset, imo with correct settings, can provide better treatment and reduced AHI

[SleepWrangler]

Does anybody know how the ResMed S9 machines decide an apnea is a cenral rather than an obstructive?

Obviously it has no input to"see" if chest is "trying" to breathe--so how does it decide that a central is happening?

This document may answer some of your questions:

Central Sleep Apnea Detection and the Enhanced AutoSet Algorithm
http://www.resmed.com/fr/assets/document...-paper.pdf