What does this tell me Sleepyhead

What does this tell me Sleepyhead - Printable Version

+- Apnea Board Forum - CPAP | Sleep Apnea (https://www.apneaboard.com/forums)
+-- Forum: Public Area (https://www.apneaboard.com/forums/Forum-Public-Area)
+--- Forum: Main Apnea Board Forum (https://www.apneaboard.com/forums/Forum-Main-Apnea-Board-Forum)
+--- Thread: What does this tell me Sleepyhead (/Thread-What-does-this-tell-me-Sleepyhead)

Pages: 1 2 3 4 5

RE: What does this tell me Sleepyhead - TASmart - 05-13-2017

I use both the P10 and an Eson 2 nasal mask, current using the nasal mask as selected on the machine. I am thinking that the machine is off calibration on its leak rate calculation. I also think, 9don't know) that the DME will not think this is a serious issue, I will contact them on Monday to see if they have anything useful to say. AT one point because I did not like the leak rate that was shown I looked to see if the vent leak rate and slope were settable, could not find anything. Which of the three setting assumes the highest vent leak rate? I glean that its the full face setting? If so I am happy to try that setting, although right now I am tending back to a fixed pressure and to stop messing about, But now I am curious.

Incidentally, my machine only has three mask setting, nasal, pillow and full face. The implication there is that in the absence of large leaks the accuracy in unaffected.

RE: What does this tell me Sleepyhead - TBMx - 05-13-2017

(05-13-2017, 11:27 AM)TASmart Wrote: The implication there is that in the absence of large leaks the accuracy in unaffected.

Well ... that's a really wild guess.

xPAP-machines have only 2 sensors - the pressure-sensor and the flow-sensor. (although I tend to believe that some devices don't actually have a real flow-sensor as the flow can also be calculated (pretty accurate) with the pressure and the already known speed of the turbine)

the measured flow is the total-flow .. meaning: respiratory flow + Co2-Vent-flow (intentional leak) + real leak (unintentional)
intentional leak is not that complex and can be estimated ... the hard part is telling the real leak from the respiratory flow apart.
In the end they use the pressure and the flowrate to measure resistance or conductance and calculate the leak from that.

whatever remains is taken as respiratory flow - if they miscalculate the leak - the flowrate is also false.

as I said: I really don't see there a hypopnea ... but if you shift your (real) respiratory curve 6 liters up and only count the upper half (your inspiration) as tidal volume you might end up seeing hypopneas where you only got the leakage wrong.

the mask-settings should preset this resistance .. as I don't have a resmed-device I cannot comment on the order of the 3.
So I would try pillow one night and see if it makes a difference.
(are there any noticeable differences in the leak-rate between the two masks - both with "nasal" setting?)

RE: What does this tell me Sleepyhead - PaytonA - 05-13-2017

Hypopneas come in 2 different flavors, obstructive and central. If your hypopneas were central then the machine will not raise pressure based on the hypopneas. You had primarily CAs and hypopneas during the time that your pressure was low. When you finally had one OA, the machine did raise the pressure. In addition, Your flow limitations did reduce significantly as bonjour said.

Best Regards,

PaytonA

RE: What does this tell me Sleepyhead - TBMx - 05-13-2017

Resmed does not differentiate betwenn obstructive and non-obstructive with hypopneas (otherwise you would see the FOT in action on the last high-res screen)

.. but resmed (only) cares for flowlimits - if non is present there is nothing it will do.

RE: What does this tell me Sleepyhead - TASmart - 05-13-2017

I just checked in the manual, the tolerance in the leak rate is +/- 12, therefore any leak of less than 12 is unreliable. It supports that I have no leakage the majority of the time within the ability of the instrument to detect.

I have never checked the reported leak rates with different masks attached. I limit my use of Pillows because i get a higher AHI with the pillows than with the Nasal.

I wonder if this means that my AHI is to some degree miscalculated and the pillow are really not worse AHI wise then the nasal mask?

RE: What does this tell me Sleepyhead - TBMx - 05-13-2017

(05-13-2017, 12:27 PM)TASmart Wrote: [..]therefore any leak of less than 12 is unreliable.

Your conclusion is wrong!

(That's why I am such a leak-"nazi")

EACH AND EVERY leak can be off by 12 Liters per minute.

this means: if it shows zero it can be 12 - if it shows 13 it can be everything between 1 and 25.

If you take the same Screenshot as befor and rescale the flow-rate to -40 to +40 you should see, that it gets the baseline wrong! (I would bet on sth about 6 liters per minute^^)

and if you're already on it, please switch flowlimit with tidal-volume ... but that's just my own curiosity if the tidal volume really is calculated just from the inspiration on resmed.

RE: What does this tell me Sleepyhead - TASmart - 05-13-2017

I cannot agree that my conclusion was poor, my explanation was. There cannot be a negative leak rate, unless the system is being inflated either by the fan or by a persons exhalation, both are limited by the Co2 venting. So any value less than 12 ranges between o and the reported value +12. Thus it cannot be stated to truly be a leak. Any value over 12 can be said to be a leak with a value of the reported value +/- 12.

What the reported leak rate's true value also is dependent on the precision of the value. If the value is simple offset but highly precise, then it's pretty easy to say a value of 12 is actually 0. Alternatively, if the value lacks precision, the any particular value is in that +/- 12 range. I suspect the reading are more repeatable than that, but I really do not know.

That said, I am not sure what you are saying here. rescaling the flow to -40 to + 4- still gives a mean of 0.

Anyway, all that in the weeds stuff aside, here is the revised plots you asked for.

RE: What does this tell me Sleepyhead - PaytonA - 05-13-2017

(05-13-2017, 12:17 PM)TBMx Wrote: Resmed does not differentiate betwenn obstructive and non-obstructive with hypopneas (otherwise you would see the FOT in action on the last high-res screen)

.. but resmed (only) cares for flowlimits - if non is present there is nothing it will do.

Resmed does not discriminate obstructive and non-obstructive hypopneas in their reporting because it is not necessary since the obstructive types will be accompanied by flow limitations so why not just key off the flow limitations which is what you said. One would not see the FOT in action for a hypopnea because there is no flow cessation to key the FOT to initiate.

My thoughts.

Best Regards,

PaytonA

RE: What does this tell me Sleepyhead - PaytonA - 05-13-2017

(05-13-2017, 12:27 PM)TASmart Wrote: I just checked in the manual, the tolerance in the leak rate is +/- 12, therefore any leak of less than 12 is unreliable. It supports that I have no leakage the majority of the time within the ability of the instrument to detect.

I have never checked the reported leak rates with different masks attached. I limit my use of Pillows because i get a higher AHI with the pillows than with the Nasal.

I wonder if this means that my AHI is to some degree miscalculated and the pillow are really not worse AHI wise then the nasal mask?

The clinicians manual for my Resmed S9 VPAP Auto does not give a tolerance for the leak rate. It has a note that says,"The displayed values are estimates. These are provided for trending purposes only."

Best Regards,

PaytonA

RE: What does this tell me Sleepyhead - TBMx - 05-14-2017

@TASmart:
sorry for that ... I'm from germany, so discussing all this geek-techno-stuff is really giving me a hard time ... sometimes I simply cannot express what I want to say in english because I just don't know the right terms / words ...

The error (estimation) of the leakrate comes from the sensor-error of the pressure-sensor and the flow-sensor. (the mathematically part of the calculation for the leak itself has no error - BUT: if you multiply values with error so does the error)

there is no sensor just for the respiratory flow. We obviously do trust that! Event-detection / the flowlimit all the fancy stuff ... it all really comes from the respiratory flow-curve.
If you do not trust the leak there is no reason to trust the flow either!

once again: 1. the machine measures the pressure and the total flow - both with measurement errors.
2. than the device does some fancy calculation to get the total leak with the pressure and total flow
--> Leak-flow = F(pressure o total-flow)
depending on the function F the error changes accordingly.

3. the respiratory flow than is: Respiratory-flow = total-flow - leak-Flow <--- whatever is the result here, is the input for the event-detection and the response of the Apap-algorithm.

4. the unintentional leak is than: unintentional-leak = Leak-flow - intentional-leak.
(usually all machines somehow get the intentional-leak pretty accurate ... but that seems to be more or less equal between all masks of the same type and changes more over the mask-types.)

In the function F in step 2 they do calculate resistance or conductance as one step in the calculation .. so whatever you choose in the settings has an impact here! The 'calculation' is a little bit more than just a sum or a simple multiplication ... so if you have an error in the raw-values the error increases accordingly.
This calculation is not done once for a breath or simply to show you a leak-rate but for each and every data-point to get the respiratory-flowrate out of it and do the *magic* with that very result.

The pressure in the mask is not that constant as one could believe from the pressure-chart. If you inhale to take air out of the system - resulting in an instant pressure-drop in the mask. The device does not measure at the mask - it measures at some point in the device itself - long before the tube even starts. If you exhale you put air in the system resulting in an instant increase in pressure. These pressure changes are not instantly distributed throughout the system - this takes some time (although measured against hard reality that goes pretty fast^^)
The devices need to collect the flowlimitation or obstruction which is - in the end - nothing more than your actual flow-curve compared to a perfect breath (or the difference to a "sinus-curve"). If the device - somehow - gets just some data-points during one breath "wrong" you are way off from a perfect breath - thus you see flow-limitation. (once again: the momentary pressure goes into that calculation as well!)

Now one can come to the conclusion that it is all crap and everything the machine does is simple guessing as everything is prone to error.

As I said: I'm a data-believer!

All the fancy stuff the machines do collect have some error - that's true. But the error from the leak-rate of +/-12 liter per minute is measured against the hard and absolute reality! It is of no importance for the treatment if the machine always gets the pressure wrong by a few mmH2O or always is some liters per minute off with the total flow.
What is of importance is only the change in these errors! (or the change of these values - thus the (absolute) error is not that big)

real-world example: if you want to measure the area / size of a football field and you only have an "inch rule" (I hope I got this right) of 1 meter length, which can measure the length with 0,1% error or 1 mm per meter. The amount of squaremeter you will measure is - sompared against the hard reality - crap!
But if you do want to compare the size of 2 different football-fields you will be able to say pretty accurate how much they differ (in percentage)!
That is the very same thing the machines do! - all that matters is the difference or the course (over time) of the charts / data.

If you do put some trust in the course of the pressure-curve AND some trust in the course of the flow-curve (or for that matter simply everything the machine does and shows you!) than you can also believe the leak-rate!

I always yell over the leak-rate because - at least for me - it is the easiest way to see how much error is included in all the data presented. - So, I simply use that as an indicator how reliable all the other things are! (and not just - as in your case - generally but for every given data point I do that evaluation)

If we take the timeframe around the hypopnea in the 2-minute-shots ... you have an average breathing volume there about 6 liters per minute (as measured by your machine). As you said the leak can be off by 12 liters per minute (and so can be the respiratory flow) ... so what do we take out of that? you could have stopped breahting totally? - your breathing was around 18 liters per minute? - which would mean over 1 liter per breath - pure AWESOME!
All I would take out of these values (if I would trust them) is: your breath got more shallow - roughly to the same amount as during the RE-part of the RERA ... after quite some time you made some recovery breath and your breathing returned to "average". If you really took 389 ml per breath or 321 ml or 456 ml is of no importance at all!

As I said: "If I would trust them!" - which I don't in your case.

As the values from the 2 sensors go multiple times in the calculation of the leak, the error increases. One can see that very same error also in the flow-rate ... but it is a bit harder to see there as the scaling is tooo high and the flow - of course - changes drastically over time. So it's easier to just look at the leakrate as it is more "even" and better scaled by deafult.

As you pointed out there is between each breath a point with zero respiratory flow. In the flow-charts these points should be located on the 0-line. (not under not over)
I tried to mark them on your last screenshot (the timeframe during this RERA and the Hypopnea is obviuosly not ideal for that - but you will see what I mean if you look on a time with no events and evenly breathing)

Clearly your machine got the leak-rate wrong and therefore got the flow wrong - all the time - not just there!
So what's the point in looking any further at already false data to speculate on what might or might not have been there?

And that is the very same data your machine operates and calculates on!

That's why I always say: if you are on any kind of "auto"-pap you have to get your leak-rate under control! Anything and I really mean anything around leaks (not just if it flags it as leaks - some (little) time before and after is already falsified as well!) is a very rough guess by the machine.
There is no such thing as an accceptable leakrate on "auto" ... these numbers only hold for CPAP! (as there is nothing the machine has to interprete or get out of the data - except keeping the pressure steady!)

Increasing the pressure always reduces the error! (as shown above) ... but that should not be an end to the means! - get the leak under control first - after that - with really reliable data - do the adjustments - you really get much more out of it!
(Or in your case - make the presets right)

I'm not sure if I could make myself as clear as I want to.
I do not say that the machines are unreliable - I just say that you really have to be careful with your own interpretation of the data. The machine does not tell you that there is (or was) an apnea. It just tells you, that to its own findings it came to the conclusion that the respiratory flow ceased or dropped below 25% of the mean average of the last timeframe ... and if it makes a distiction between clear airway and obstructive it does not state obstructive or central apneas - it simply tells you: "if I increased the pressure there I could detect some flow" - nothing more.
I do check first if I personally have made everything right so that the machine can measure "correctly" .. some things I can control - some are simply out of my own control.

@PaytonA:
you are - of course - absolutely right with the hypopnea distinction. As there still is a flow-curve, one can get the level of obstruction or flowlimitation out of that. Introducing a FOT during the hypopnea would yield more precise results - but most likely wake the patient or at least cause discomfort.
I have to admit that I am generally over sarcastic against ResMeds-Hypopneas (and RERAs) ... but that's mainly because they don't score a length for both and all it tells you is in the end: there was sth somewhere around that point ... go figure it out yourself^^ (well ... for me the tidal-volume holds much more information than that Hypopnea-flagging does!)

I have to admit I'm really unsure if there is a device out there which actually would start sth. like FOT on a hypopnea and I also tend to believe that my machine (H+L) makes the distinction of the hypopneas solely based on the flow-curve and the presence or absence of obstruction. (but I do not know for sure so far)

but in the end this is - once again - a problem with the interpretation (in this case mine) and not so much the presentation or the data itself.
(and hypopnea detection is much more prone to errors .. you have less flow - therefore a greater impact of the measurement error ... and after that there is the "thingy" with the "compared to mean average" ... whatever you choose as that timeframe changes the outcome - if you took for some time very deep breaths every normal breath after that will be a hypopnea by definition.)