Pressure Relief Features (JCSM article)

[DaveCar]

Posted in the Journal of Clinical Sleep Medicine in 2016.
If this has been posted and discussed already, my apologies. 

Summary:
"Pressure-relief features may lead to attenuated CPAP treatment efficacy depending on the applied settings and the device. In clinical practice, the therapy efficacy can be ensured by increasing the therapeutic pressure or by enabling the pressurerelief features prior to the manual or auto titration process. The pressures in the pressure-relief APAP device reports are not comparable to that of conventional APAPs."

https://jcsm.aasm.org/doi/pdf/10.5664/jcsm.5590

For APAP users, would it be fair to say that a larger exhale relief setting should also be accompanied by a corresponding increase in the min-PAP setting ?   ie - going from EPR (or Flex) 1 to 3, then also change min from 9 to 11. 

I tried P-flex on my new DS2.. thats some crazy pressure variations, couldnt get to sleep.  A-flex works fine though.

[Sleeprider]

Never was a fan of the Philips "Flex" pressure relief. It is disruptive and the Philips algorithm for auto CPAP does not increase pressure and maintain it enough to stop many events. Fixed CPAP with pressure relief is a bad idea for most people that do not understand how to mitigate for obstructive events when using pressure relief. The Resmed EPR produces a true bilevel pressure, where the exhale pressure needs to be titrated to avoid obstruction. We see many new members on this forum present their first chart with minimum pressure at 4.0 and EPR at 3, which means the the auto algorithm is challenged with getting pressure above 7.0 before obstruction can even begin to be addressed. Fortunately Resmed detects flow limitation and responds faster with pressure increases. Pressure relief has its proper place in comfort and even therapy, but it takes either good coaching or a solid understanding of how things work to not lose efficacy. There is over-reliance on auto CPAP and the use of 4-20 pressure, which is a bigger problem than the pressure relief itself.

[DaveCar]

Re the 4-20 lazy-set
Do any machines 'learn' a therapeutic level and then apply it each night thereafter ?

[clownbell]

As I understand it, the machines do not "learn" anything to apply in future. They react anew each night to the variables presented to them.

BTW, I really wish these studies could be written in a manner which a reasonable person could understand - not in scientific jargon which only the insiders can decipher. Then again, maybe the authors don't want the "great unwashed masses" to understand.

[Dog Slobber]

Do any machines 'learn' a therapeutic level and then apply it each night thereafter ?

No.

The one exception I'm aware of;  Respironics DreamStation's OptiStart mode.  OptiStart will start the device at the 90% (maybe 95%, can't remember) pressure of the previous nights pressure. 

I'm not sure that I would classify it into the category of learning or even optimizing.  But it does fit into the description of basing the current session on previous sessions performance. 

I don't trust it, and I don't believe it is used much.

[SevereApnea]

@DaveCar

"Do any machines 'learn' a therapeutic level and then apply it each night thereafter ?"

Nope, not yet, not sure why the manufacturers haven't jumped on the AI bandwagon yet, it is all over the rest of the world, in IT, in computers, medicine even Webcams etc etc. 

Here is a post I made yonks ago, which had no replies, so maybe we as users are not ready for this yet ... FoodForThought

Maybe the Lawyers have put the brakes on? but anyway we should be able to Opt In or Opt Out?

Maybe with the Philips FoamGate saga the pressure is to just get as many machines out there as fast as possible rather than develop new products?


I think this is still in Science Fiction mode but maybe some manufacturer developers are hard at work at this atm.

Here is a more recent article exploring the BARRIERS

Sweet dreams!

[DaveCar]

I suppose, in a time of liability, AI problems would be the supplier's responsibility.
In a clinic/doctor led config, mistakes would be the clinic/doctor's responsibility.

[SleepyHenry2]

I feel uneasy when I see bench “laws” of physics applied to biological systems, as in the above article.  However, I read the article and concentrated on EPR.
Their Outcome Measures are residual AHI and AI.
Their research protocol measures waveforms produced by laboratory produced Obstructive Apnea.
I quote from their results: ”This modality of pressure delivery was not efficient to maintain the airway patency when apnea occurred (Fig 1B, columns 2 &3).  For EPR the therapeutic pressure decreased by 3 cm H2O and apneas thus persisted (Fig 1B,column 4).
A lemma of this forum is EPR=3 is for hypopnea and flow limitations.... So nothing new in their findings about using EPR.

What I like to bring up for discussion is their conclusion that the pressures in the pressure -relief APAP device reports are not comparable to that of conventional APAPs (Table III)?.... Not sure if this is conventional wisdom?

[StratCat48]

@DaveCar.........

"I tried P-flex on my new DS2.. thats some crazy pressure variations, couldnt get to sleep.  A-flex works fine though."

My DS2 only has 'Flex' 1,2,3 settings. Has the P-Flex and A-Flex been an addition to the DS2 with a recent firmware/software update? My current version is V1.0.3.3690.

[Gideon]

Conventional APAP means no EPR.
This when you apply EPR = 3 you drop EPAP by 3 cmw which assuming the pressure was properly titrated originally ePAP pressure will drop below the optimum pressure to treat OA events, BUT that is only if you don't adjust pressure to compensate.
Often the "pros" here don't knowing that in most cases the original settings were too high. They do, on the next set of charts, look to see if this happened and will adjust them.

EPR is not a negative but rather a positive that treats many types obstructive events