C-Flex and the I:E ratio

[FrederickRose]

This could be considered an extension of an earlier thread: viewtopic.php?f=1&t=113290&st=0&sk=t&sd=a&start=90

Please merge if applicable.

I noticed that my PR DreamStation with C-Flex 3 gives me an "inverse" (greater than 1) I:E ratio:



Whereas my ResMed with EPR 2 gives me a "normal" (less than 1) I:E ratio:



Ideally I would have used CPAP for the comparison as opposed to APAP which gave me different pressures at different times, but the I:E reading is for the entire night, and my APAP pressures were similar for the two nights.

I then tried PR DreamStation with C-Flex off, and that gives me a normal (less than 1) I:E ratio:



My ResMed with EPR off also gives me a "normal" (less than 1) I:E ratio:




Unanswered questions:

1)Are these I:E ratios correct, or is the PR measuring the start of inspiration prematurely when C-Flex is enabled. You can see that there is a plateau of minimal flow preceding increased inspiratory flow, and the machine seems to be marking the beginning of inspiration at the beginning of that plateau

2If they are correct, does it matter, or is this all academic?

I'm guessing that the answer to #1 is that it is a faulty measurement. Why would C-Flex cause the first part of inspiration to be stalled at a low flow rate for a while before starting in earnest? I'm thinking that inspiration should actually be measured from the end of that plateau.

[Matt00926]

Just a wild guess, but maybe with more exhale relief, it's easier for you to take a deeper (which ends up being longer) breath.

I think it's harmless either way, as long as you don't suffer from an obstructive lung disease like COPD. And the I:E ratio is based on your spontaneous breathing, as opposed to being specifically set via mechanical ventilation.

[Jay Aitchsee]

Hi Fred,
First, realize the I:E ratio has no diagnostic value. It is a metric that applies to ventilation only, that is, when a machine is breathing or assisting in breathing for the patient. In ventilation applications, the I:E ratio is set to deliver the appropriately timed breaths to provide the desired blood gases. So in reality, it is a setting rather than a reading or measurement.

SleepyHead does provide the I:E ratio since the measurement is available from some machines. Generally those machines, advanced models of which, are ones which may provide some ventilation. This includes many Bi Level lines of machines. The I:E ratio is generally not available from machines which have no ventilator capabilities, i.e., autopap and cpap.

In the thread you referenced, viewtopic.php?f=1&t=113290&st=0&sk=t&sd=a&start=90, a lot of study of the derived I:E ratios was done. My determination was that, generally for PR machines, the derived I:E ratio was in error because expiration time was stopped when a ballistocardiographic "bump" was encountered (as you noted in the Cflex example). Ballistocardiographic bumps are those little bumps which occur during the expiratory portion of the flow wave. They are caused by the pressure differential in the lungs experienced when the heart beats. They are normal. Some conditions of xpap cause them to be more apparent than others. Apparently, Cflex is one. I think, but I am not sure, that ResMed filters the heart rate frequency from the flow so that this phenomenon is not as apparent in ResMed machine timing reports.

When the I:E ratio is not computed, as in your case and in the case of most auto and cpaps, one may compare the statistics given to derive the I:E. The machines yield min, max, average, and 95% values of inspiration and expiration. However, that would be incorrect. One can not simply divide one inspiration statistic by the corresponding expiration statistic to arrive at an I:E. These are averages over night. To be correct, one needs to compare the inspiration period of one breath to the expiration period of the same breath.

The bottom line is that if you want to know what your true I:E ratio is when using a PR machine, you must measure and compare the inspiration and expiration periods of the flow wave graphically yourself.

Even with a ResMed machine, a derived I:E value from the statistics doesn't make much sense because it is measuring and comparing average values rather than specific portions of the same breath. I think the best that can be done if I:E ratios look "wrong" is to examine the flow wave in detail as you've done and determine why it is so. In your case it is the cardiographic bumps. In others, it could be excessive mouth breathing, or something else. In any case, there is no reference, that I have found, that points to, or uses I:E as a diagnostic tool.

It should also be noted that even though derived I:E ratios from PR machines appear to be incorrect, that does not mean that PR machines can not provide ventilator functions, but In the case of cpaps, autopaps, and many bi levels, the I:E ratios have no valid application.

https://en.wikipedia.org/wiki/Ballistocardiography

[FrederickRose]

Jay, thank you for the information. So helpful!

[FrederickRose]

To be correct, one needs to compare the inspiration period of one breath to the expiration period of the same breath.

The bottom line is that if you want to know what your true I:E ratio is when using a PR machine, you must measure and compare the inspiration and expiration periods of the flow wave graphically yourself... I think the best that can be done if I:E ratios look "wrong" is to examine the flow wave in detail as you've done and determine why it is so. In your case it is the cardiographic bumps...

Visually, the I:E looks to be > 1 with PR + C-Flex:



Are you saying that those bumps in the beginning of what is shown as inspiration are actually part of expiration at that the zero flow line should be moved up?

Here is a similar segment with the ResMed, where the bumps are below the red line (in expiration) and visually the I:E looks to be < 1:

[Jay Aitchsee]

Are you saying that those bumps in the beginning of what is shown as inspiration are actually part of expiration at that the zero flow line should be moved up?

Generally, yes.

In your example, I think the zero line should be around 7. I think this could be seen better if you would look at fewer breaths on the horizontal scale, expanding it to better be able to see the difference in slope between exhale portions and inhale portions of the flow curve. I also think this will change when looking at different segments of your graph.

I am not an expert in Cflex http://c.ymcdn.com/sites/www.gameshme.o ... s_2016.pdf but it apparently changes the pressure profile on a breath by breath basis and this changing pressure apparently has an effect on the resulting display of Flow. Now, I don't know if this apparent discrepancy lies with SleepyHead or the PR machine. I suspect SleepyHead is not accounting for the effects of CFlex. My guess is the display would be different if rendered by PR Software.

But here's the thing, we are talking about a parameter (I:E) that has no valid application in basic cpap. SleepyHead is able to derive Inspiration and Expiration time from the flow rate data, but that does not mean it is correct. It might be, it might not be, depending on the circumstances. Interestingly, SleepyHead presents Inspiration and Expiration times for my S9 autoset, but ResScan (ResMed Software) does not. I believe this is because the variables are not appropriate to autoset application. ResScan does provide Inspiration Time, Expiration Time, and I:E for machines that have ventilator capabilities.

People look at the Inspiration and Expiration times as presented by SleepyHead and calculate or observe an "Inverse" I:E ratio. What they are missing is that an Inverse I:E ratio is not an abnormality, it is a technique. It is a technique used in ventilator application to hold the inspiration portion of the flow longer than the expiration portion for better oxygenation. Inverse I:E has no application or meaning in non ventilator application. https://en.wikipedia.org/wiki/Inverse_ratio_ventilation

Here's a thread with comments from some old-timers about ballistocardiographic artifacts you may find interesting:
viewtopic.php?f=1&t=61466&st=0&sk=t&sd= ... 56#p576977

[ShinRyoku]

I was posting here under the name FrederickRose. I've re-registered under the name ShinRyoku in order to standardize across a few different sleep apnea websites.

"In your example, I think the zero line should be around 7"

If we move the line there, however, it would seem that the total amount of air exhaled would greatly exceed the amount inhaled, which isn't possible. Something else would have to change, no?

"What they are missing is that an Inverse I:E ratio is not an abnormality, it is a technique. It is a technique used in ventilator application to hold the inspiration portion of the flow longer than the expiration portion for better oxygenation."

I understand that it is used for that application. But it's also true that people who are healthy and breathing spontaneously without positive pressure ventilation normally have an I:E ratio less than 1, and therefore having an I:E ratio greater than 1 is abnormal. Whether it is clinically significant is a separate issue from whether it is abnormal.

[Jay Aitchsee]

If we move the line there, however, it would seem that the total amount of air exhaled would greatly exceed the amount inhaled, which isn't possible. Something else would have to change, no?

Yes, the total volume of exhaled air should equal the total volume of inhaled air and the total volume is equal to the area under the curve--If we have a closed system, but we don't. The mask has a vent and you have indicated a nasal mask, so there is the possibility of mouth venting as well. I haven't worked with PR systems enough to understand how Cflex works and how various leaks or vents might affect the Flow Rate graph. I do know that ResMed treats mask flow rate differently than PR.

I will give you that moving the line up to 7 looks like it would result in an imbalance of inhaled vs exhaled air, but I can't be sure.

Here is an expanded view of your cflex example from above, rather hastily and sloppily done -Sorry. It is the second breath from the left, the one with the distinct shoulder on inhale. The red/black line is the original 0 line, the thin, red, hand drawn line is at the shoulder which is approximately at 7. I don't think there is any doubt that that is where inhale starts, but, as you point out, the volumes are not equal. Exhale appears larger. But if the thin red line is the 0 point, then Ti is less than Te and the ratio is not inverse.



So my next question is, So what? What are you looking to prove, solve, or find out. Since this condition (IVR) does not seem to appear with Cflex off or while using a ResMed machine, I would consider the result highly suspect when using CFlex.

I do take your point that an inverse ratio while not on a ventilator would not be "normal", but I think that most instances seen while using a cpap machine are artifacts of some sort produced by some anomaly in the therapy such as leaks, mouth breathing, or improper use of the information provided. It does seem that most who report an inverse I:E are PR users, but that is only a perception. I do think that actual inverse I:E is extremely rare. What could cause it? Hyperventilation? Most illnesses, I believe, cause I:E to become smaller not larger (inverse) and I still have not been able to find one instance of anyone using an inverse I:E ratio for diagnostic purposes or even a mention of what it might mean if an inverse I:E was experienced by someone not on a ventilator.

Edit: Curiosity got the better of me and I printed the above on a piece of graph paper and then counted the squares enclosed by the graph,
Here's what I found:
For the instance of the original: the volume was inspiration=79 squares, expiration=67 squares, Greater Vi
For the second instance with the 0 line moved up: inspiration=56, expiration=94, Greater Ve
Conclusion: ignoring any open system leaks, etc. the Zero line should be moved up, but not quite so far as I did in the example.

[ShinRyoku]

Yes, the total volume of exhaled air should equal the total volume of inhaled air and the total volume is equal to the area under the curve--If we have a closed system, but we don't. The mask has a vent and you have indicated a nasal mask, so there is the possibility of mouth venting as well...

I will give you that moving the line up to 7 looks like it would result in an imbalance of inhaled vs exhaled air, but I can't be sure.

The area under the curve looks roughly the same for I and E when C-Flex is off (see first post). So hard to understand why the line would be off AND the flow rate curve adjusting differently for leak when C-Flex is on.

So my next question is, So what? What are you looking to prove, solve, or find out. Since this condition (IVR) does not seem to appear with Cflex off or while using a ResMed machine, I would consider the result highly suspect when using CFlex.

I do take your point that an inverse ratio while not on a ventilator would not be "normal", but I think that most instances seen while using a cpap machine are artifacts of some sort produced by some anomaly in the therapy such as leaks, mouth breathing, or improper use of the information provided. For the same reason, I'd rather breathe through my nose than my mouth while using CPAP.

I mostly have just an academic curiosity here. But if it's true that C-Flex causes an I:E greater than 1 and it's not an artifact of some sort, then I'd personally be a little less to use C-Flex. Not based on any good logic and certainly not based on evidence. But to the extent possible, I'd rather keep things similar to "normal" physiology when I have the option to do so without any apparent downsides.

[Jay Aitchsee]

Well, this is fun
Sorry I missed your reply post while I was editing. I will repost the edit here:

Edit: Curiosity got the better of me and I printed the above on a piece of graph paper and then counted the squares enclosed by the graph.
Here's what I found:
For the instance of the original: the volume was inspiration=79 squares, expiration=67 squares, Greater Vi
For the second instance with the 0 line moved up: inspiration=56, expiration=94, Greater Ve
Conclusion: ignoring any open system leaks, etc. the Zero line should be moved up, but not quite so far as I did in the example.

I mostly have just an academic curiosity here.

Academic curiosity is as good a reason as any to pursue this subject! Or any other, for that matter.

I understand what you are saying about using cflex (if it's going to cause an inverse I:E), but I doubt that it is. My guess is that this is an artifact caused by using information in a manner it was not intended which could be induced by using SleepyHead rather than the manufacture's software, which is Encore, I believe (if you're interested, Pugsy can tell you how to get it. My understanding is that it is a PITA to work with, but it may be worth it in the search for knowledge

One other thing is the manner in which you are deriving your I:E. I'm guessing you're comparing the statistical averages to get it. As I said somewhere above, this may not be accurate

Here's and example, six periods of I/E: 1/2, 1/2, 1/2, 1/2, 4/1, 4/1; Average = 12/10 (inverse), but only two periods are actually inverse. A bit extreme, but you get the idea.

[Pugsy]

Encore, I believe (if you're interested, Pugsy can tell you how to get it. My understanding is that it is a PITA to work with, but it may be worth it in the search for knowledge

Encore won't offer much because it doesn't show the flow rate like SH does and it doesn't really give any I:E ratio information at all unless using one of the high end machines were that information might be more useful.
Encore doesn't allow the user to manipulate any of the graphs at all. What you see is what you get and it pales in comparison to what you can do with SH.

Let me go fire up Encore Pro and see if there is anything useful in there for what is being talked about here. I don't remember anything though.
I think I still have the data from someone with a DreamStation cpap/apap in my Encore.

[Pugsy]

There's zero mention of I:E ratios of any sort on an Encore report from a DreamStation cpap/apap machine. Not even any summary numbers.
The most you will get or be able to see is the wave form graphs for the flow rate and I don't know that they could be manipulated enough to do any sort of measuring of anything.

This is what the wave form graphs look like. There is nothing in Encore that will allow the user to zoom in or anything like that from within the software like we can with SH.

[FrederickRose]

One other thing is the manner in which you are deriving your I:E. I'm guessing you're comparing the statistical averages to get it. As I said somewhere above, this may not be accurate.

I was eyeballing random samplings like the ones I showed here, where each individual breath has an apparent I:E ratio > 1, and finding that they correlated with the ratio of the statistical averages. Not a bullet proof method but suggestive all the same.

I wonder if there is a medical science liason for Philips whom I could ask this question (whether C-Flex changes the I:E ratio) of.

[Jay Aitchsee]

There's zero mention of I:E ratios of any sort on an Encore report from a DreamStation cpap/apap machine. Not even any summary numbers.The most you will get or be able to see is the wave form graphs for the flow rate and I don't know that they could be manipulated enough to do any sort of measuring of anything.

This is what I'm thinking and I've alluded to it before: The manufactures, ResMed and Phillips, have not made any ventilator type data, such as Inspiration/expiration time, I:E ratio, Respiration Rate, Tidal Volume, etc. available to users of their non ventilation capable machines through their associated software. I believe this is because this type of information really doesn't apply to ordinary cpap application. This information could be derived from Flow Rate, which is what I'm guessing SleepyHead does.

I doubt if Phillips or ResMed would care if the SleepyHead presentation of ventilation variables associated with one of their non ventilator machines was not correct. I'm pretty sure their response would be that it wasn't their software, and besides, the data wasn't applicable to their machine's designed use, anyway.

It is possible that Phillips and ResMed apply special filtering and algorithms to take care of things like the ballistocardiographic artifacts that we've discussed earlier when deploying their machines as ventilators. In fact, I'm pretty sure I've seen patent applications from one or both discussing methods to prevent false triggering when these types of artifacts are present.

And that leads to the supposition that SleepyHead may not be able to present some ventilator type information from non-ventilator machines correctly because it may not have the necessay tools or data to do so. Just a thought.

[Jay Aitchsee]

How about this?
To eliminate the artifact:
Extend the inspiration portion down to the zero line following the same slope (orange line).
Now, I:E becomes the green line over the purple line which produces a "normal", less than 1, I:E



It's my opinion that your actual I:E is not inverse when using Cflex. I think it's only the graphical representation coupled with false measurements by either SleepyHead or the machine which makes it appear so when cardiographic artifacts are present.

In the other thread, posting.php?mode=quote&f=1&p=1093893, on the same subject, the implied conclusion was that PR machines simply did not provide sufficiently detailed data. Here's an exerpt:

The thing that gets me is, since P Resperonics must have built in the means to measure event time very accurately, if for no other reason than to honor the 10 second rule - no surprise considering it's recording data points 125 times/second - why it can't report duration accurately. (Could Mark be wrong about the "markers?")

the machine may sample the flow waveform 125 times a second, but that doesn't mean it *records* data points that often.

resmed records data points 25 times a second, as to how often they sample the sensors? I don't know off hand.

respironics data recording has always been less precise than resmeds, just look at the pressure traces that are displayed, resmed steps pressure in minute amounts, respironics seems to step it in halfcm amounts, and then there's the mask pressure trace, which is *measured* pressure on the resmed, the only respironics to have that data is the ASV machine (iirc).

as to marks guesses (because he has no official documentation)... who knows. they're not an unreasonable guess, but guess or not, he can't display data that's simply not recorded.