Inspiration expiration ratio

[rkl122]

..............However, Sleeprider and I agree, if an Inverse I:E is present, it might suggest an abnormality. However, neither of us has been able to find a reference where it was used as such.

Understood about the diagnostic issues. But guess I'm still a little confused because SleepRider presented the data of an AHI=20.57 person as representing the abnormality we're discussing, but the (95%) I:E ratio is less than unity (1.96:4.88) - ie. not "inverted". Are you guys actually talking about that next line up that reads "I:E Ratio"? What is that? Apparently it appears only in Resmed machines - my Respironics machine doesn't report anything like that. I've been referring only to the Inspiration:Expiration ratio as represented by the ratio of the 95% inspiration : 95% expiration values, given the fact that the corresponding median and low ratios are also inverted.

I don't doubt that your nightly statistical I:E Max is greater than one. I would be surprised, however, if your statistical average for an entire night was greater than 1.0 as you imply when you say your I:E, "...is greater than unity, as normal".

Now I'm confused as to what you mean by "statistical average". Look again at the first diagram I inserted in my second post above. The min, median, 95%, and max ratios are all greater than unity. "Inverse" as you would say. What am I missing here? Am I wrong to assume that the statistical average is represented by those figures? In the passage containing my "greater than unity" reference, I was questioning why SleepRider referred to my ratio as "normal," since my ratios are inverted by your definition of that term. But maybe I misunderstand what he is saying.

Do I need statistics 101?

-Ron

[palerider]

It might help if you adjust the trigger and/or cycle sensitivities. If it were me, I would start by lowering the sensitivity on the trigger a notch and see what happens to my flow rate curve and tidal volumes.

Excellent suggestion, Wicked, but is his PR System One REMStar 60 Auto CPAP, capable of trigger and cycle sensitivity adjustment?

nope.

[rkl122]

It might help if you adjust the trigger and/or cycle sensitivities. If it were me, I would start by lowering the sensitivity on the trigger a notch and see what happens to my flow rate curve and tidal volumes.

Excellent suggestion, Wicked, but is his PR System One REMStar 60 Auto CPAP, capable of trigger and cycle sensitivity adjustment?

No, never seen hide nor hair of them. Don't know what they are, but I want 'em.

I think that the machine is recording 2 tidal volumes for each actual breath you are taking.

Thanks for the observation WickedLoki. In a one-minute window such as the last image I posted, I notice on careful inspection that there are subtle inflection points in the tidal volume graph. So far, they appear to occur only when the flow graph meets the zero flow line, but maybe I'll be able to confirm or refute your hypothesis. BTW, what is the CPAP definition of a single breath? I've been assuming it begins the instant an inspiration leaves the zero line and ends when an expiration first touches the zero line. Which means, of course, there's a crossing in the middle of the breath.

BTW, [he says, risking more microcosmic granularity], I get the impression from the literature that there's a clinical distinction between the terms "expiration" and "exhalation". I think the former ends when the breath ends (as just defined), but if there is an extension of the expiration along the zero flow line, the latter includes that and ends at the instant of the next inspiration. I mention this because there was reference above to the possibility that our machines may get confused when there's a "tail" along the zero line.

-Ron

[WickedLoki]

His Sleepyhead data page says that what is being used is an Aircurve 10 Vauto in Vauto mode!!!

[WickedLoki]

Sorry, reading the wrong person's charts. Sleeprider, you might benefit from the adjustment mentioned.

[Jay Aitchsee]

Do I need statistics 101?

Sorry Ron, I neglected the fact that you had posted that your averages for the night were greater than 1.0, and I agree that isn't normal or typical.

What it tells me is that the shape of your flow waves are not typical. There has been a little confusion in this thread about whose posted waves are whose, but I don't think we've seen good examples of yours. Say 1 minute in duration which would show about 12 complete cycles. If you could post some, maybe we could discuss yours.

I think we'll see that the expiratory portion of your flow wave will not be typical, as both SR WL have alluded to.

Wicked did bring up a good point regarding trigger and cycle settings (and now we're getting into machines that have some actual ventilation capability). Trigger and Cycle sensitivity along with Inspiration and Expiration time Min and Max setting will influence I:E. Typically, the control available will vary with the machine complexity, starting with a some with BIPAPs to a lot with ASVs.

I mentioned earlier that I had assisted someone that presented with a I:E similar to yours. He had a BIPAP with some cycle sensitivity and limited I & E times control. One of the first things that helped in his case was to reset the adjustments back toward mid range. I had forgotten that until WR mentioned settings.

Unfortunately, as we've learned, your machine doesn't have this capability, but perhaps we can continue to build a case to get you one that does.

And to continue your mention of expiration and exhalation: It is my impression the terms expiratory, expiration, etc, are most often used in a clinical setting and are often used to refer to characteristics which could be graphically presented while the terms exhale and exhalation refer to the act itself. I think inspiration ends and expiration begins at 0 where the curve crosses the x axis. Generally, auto machines Trigger (Start IPAP) after the flow rate curve passes 0 and turns positive, i.e. the patient has initiated a breath and cycle (start EPAP) when the flow has decreased to a certain level above zero. In some machines, the level of flow required to initiate IPAP or EPAP can be adjusted by sensitivity settings.

Something that occurs to me while writing this is that both the IPAP and EPAP periods would normally start during the Inspiration phase of the cycle, that is when the flow is positive and the length of these periods would be further determined by whatever I & E Time Minimum and Maximum settings that were being applied. This could offer a clue as to the results that are sometimes seen.

[rkl122]

..............I don't think we've seen good examples of yours. Say 1 minute in duration which would show about 12 complete cycles. If you could post some, maybe we could discuss yours.

Thank you Jay for this offer! I'll need a day or so to collect thoughts and images. Will bump this when ready.............

And to continue your mention of expiration and exhalation: It is my impression the terms expiratory, expiration, etc, are most often used in a clinical setting and are often used to refer to characteristics which could be graphically presented while the terms exhale and exhalation refer to the act itself. I think inspiration ends and expiration begins at 0 where the curve crosses the x axis. Generally, auto machines Trigger (Start IPAP) after the flow rate curve passes 0 and turns positive, i.e. the patient has initiated a breath and cycle (start EPAP) when the flow has decreased to a certain level above zero. In some machines, the level of flow required to initiate IPAP or EPAP can be adjusted by sensitivity settings.

Something that occurs to me while writing this is that both the IPAP and EPAP periods would normally start during the Inspiration phase of the cycle, that is when the flow is positive and the length of these periods would be further determined by whatever I & E Time Minimum and Maximum settings that were being applied. This could offer a clue as to the results that are sometimes seen.

This reminds me that I've been using AFLEX (the Respironics flavor of expiratory pressure profiles) all along. Mostly at setting 2, but for the last few nights at 3. Would that make a difference in your analysis? Let me know if you feel I should turn it off before gathering images.

Incidentally, by sliding the cursor slowly across my graphs, I just discovered something interesting. In the first hour or so of sleep, the I:E ratio is generally NOT inverted. Inversion predominates in the early morning hours, during which time these oddball events tend to accumulate as well. (In accord with SleepRider's observations perhaps.) This may be a clue? Will summarize when I post images.

Many thanks,

-Ron (CSI SleepyLand)

[Jay Aitchsee]

This reminds me that I've been using AFLEX (the Respironics flavor of expiratory pressure profiles) all along. Mostly at setting 2, but for the last few nights at 3. Would that make a difference in your analysis? Let me know if you feel I should turn it off before gathering images.

I think whatever AFLEX Setting you are using will be fine to start. AFlex works a little differently than EPR which is what I'm most familiar with. I think at some point it might be prudent to gather a few representative samples from each setting. Probably the most telling would be compare OFF to MAX ( 3?).

It is not unusual to experience more disturbances, or "events," in the early hours. Usually this is attributed to increased frequency of REM periods, as well as drifting in and out of sleep (which also could be attributed to the ending of REM stages). In these hours, a lot of so called Sleep-Wake-Junk is created. Defined primarily as CA resulting from arousals.

Now, arousals not related to Sleep Disordered Breathing are normal in sleep. The average person has many more than commonly thought. Most are not remembered. Only those that result in awakenings of some length are typically remembered. Some of these arousals could result in a disturbance which temporarily disturbs breathing and results in a CA. So, some CA would be normal. When does the number become abnormal? I don't know. However, if sleep is non restorative, than I think more than a "few" would be suspect.

Additionally, just lying in bed awake and masked up is likely to create a lot of CA and some ugly ones, too. So, you must make sure the data your looking at is when you are truly sleeping or trying to. This would include the normal arousals experienced prior to fully wakening, but not that after you have become fully awake and say to yourself, "I'm just going to lie here 5 more minutes."

The point of all this is that any disruption of the normal respiratory flow where changes in breathing are experienced such as after an arousal or while awake is likely to create a condition of "Inverse I:E", after all, in simple terms, it is nothing more than a string of breathing cycles with shortened expiratory periods (or lengthened inspiratory periods, or both).

My guess is many people have periods which would create periods of Inverse I:E, but their machines just don't report it, my latest model S9 Autoset does not. If more machines reported it, I'm sure we would see a lot more questions about it. That being said, however, I don't think we would see many reports of "Inverse I:E" which last the whole night as yours appear to. Once again, I think "Inverse I:E" could serve as good alert that there might be a problem, but it will take some investigation to determine what the actual problem and cause is.

This article is an interesting read and it has a nifty diagram showing the effects of the different types of breathing on the expiratory portion of the Flow Wave:
http://jap.physiology.org/content/105/3/854

[WickedLoki]

I do not see an I:E ratio of 1:1.3-1.4 being all that abnormal. My I:E ratio is generally in that area and Resmed's charts make it seem like it is not all that unusual or do I have this backwards?

[Jay Aitchsee]

I do not see an I:E ratio of 1:1.3-1.4 being all that abnormal. My I:E ratio is generally in that area and Resmed's charts make it seem like it is not all that unusual or do I have this backwards?

Yours expressed as decimal is less than one and would be "normal"
The op's expressed as a decimal is greater than 1, at least as I understand his posting.
Yours 1/1.3 = 0.76 compared to his = 1.4

[Jay Aitchsee]

Understood about the diagnostic issues. But guess I'm still a little confused because SleepRider presented the data of an AHI=20.57 person as representing the abnormality we're discussing, but the (95%) I:E ratio is less than unity (1.96:4.88) - ie. not "inverted". Are you guys actually talking about that next line up that reads "I:E Ratio"? What is that? Apparently it appears only in Resmed machines - my Respironics machine doesn't report anything like that. I've been referring only to the Inspiration:Expiration ratio as represented by the ratio of the 95% inspiration : 95% expiration values, given the fact that the corresponding median and low ratios are also inverted.

Ron, WL's post above made me go back and check your posts. I had assumed your I:E ration was greater than 1 throughout the night because you said it was. From that, I thought it was a machine calculation. Yes, more advanced machines that have some ventilator properties do calculate I:E and it is displayed in the statistical data - in answer to your question of, " I:E, what's that?", above.

Now I see in reading more carefully, that you are calculating your I:E from overnight statistics. I don't think that is valid and here's why: I:E is a per breath measure. You have to compare the inspiration period of a cycle to expiration period of the same cycle. Your respiration rate will vary through out the night and therefore the I&E time periods will also vary, such that you will have a time of expiration max and min as well as a time of inspiration max and min. Using your methodology, it follows that you could divide Imax by Emax and come up with an approximate I:E - you can't because they don't happen at the same time. Perhaps closer in time would be Imax/Emin or Imin/Emax, but those, no doubt, would result in wildly different numbers. I said I would be surprised if your I:E was greater than one for the whole night and I was. Now, I again have doubts that it truly is. But then, I could be wrong.

Either way, it doesn't change the fact that you still have some problems and examining waveforms and talking about I:E may help understand and solve those problems, so let's press on.

[rkl122]

Jay, thank you for spending time on this. I've been thinking along the same lines, and believe you're correct. I apologize for the erroneous logic. However, guess what - SleepyHead permits the substitution of "average" and "weighted average" (whatever that is) for median values! (Look in preferences under the CPAP tab.) When I use either of them, the numbers change a bit, but the observation of inversion remains. Hope you agree this is a more valid observation. You can see how the values compare in the link below.

.......................
Either way, it doesn't change the fact that you still have some problems and examining waveforms and talking about I:E may help understand and solve those problems, so let's press on.

Agreed, I think this is what's important now, because the more I study the flow trace at strong zoom level, the more scared I get. So let me just start with a couple windows from last night from regions where there was no nearby event, and where I'm pretty sure I was sleeping. The first is from early in the sleep session, the second is near the end. Go easy on me, but what do you think? (At least I'm breathing... )

http://imgur.com/a/ONWio Both images in one link. In the second I substituted averages for the median values for comparison sake. (Apologies for the popups blocking some data - just noticed them now.)

Thanks Jay.

-Ron

[Jay Aitchsee]

because the more I study the flow trace at strong zoom level, the more scared I get

Ron... Chill!

There's nothing to be scared about. You're letting your inquiring mind and imagination run away with you.

You have poor sleep. Lot's of causes for that. Sleep Disordered Breathing is one, caffeine is another, I'm just sayin'. Obsessing over it is a major cause - again, just sayin'.

Yes, I like average better than median most of the time. Statisticians like median, I guess, because it is not affected so much by any large outliers in the sample. On the hand, the median doesn't work well if there is a large spread in the data. Take my leaks for example, the median is most often is Zero, which doesn't tell me much, while the average might be a low number which I can compare with another night's. I like graphic presentations best where I can see what's happening.

Anyway, back to you. You have some Flow Waves that are not "typical". So, we need to investigate why that might be. The first thing is that disturbed sleep from any cause and being awake will produce waveforms that are not typical. The second thing is that we are working with inadequate and incomplete sensing capability with marginal analyzing ability. When we see something in our data, sometimes the best we can do is say, "Oh, that looks, like it might be this". The this being something that came from a lab with a full array of sensors, analytical computers, and experts who have studied the subject for years.

Sleep Labs will likely have at least 12 channels of information to analyze. We have 1, maybe 3, if we have a pulse-Ox.

Our one sensor is a flow rate detector - in the machine! It's not even close to our respiratory tract. The lab will use sensors at the nose and mouth to determine nasal and oral flow. We have no way to directly determine either. If we are using a nasal mask, any mouth expiratory flow is not detected and is assumed to be a leak equal to the sum of the inspiratory flow minus any expiratory nasal flow that is detected during that cycle. Inspiratory flow is assumed to be the change in outflow from the machine over static leak. With a nasal mask, any oral inflow is not detected. With a nasal mask, any oral inflow (mouth breathing) is not detected, not detected, and will be seen as a flat line on the flow rate graphic. Yeah, I said it twice. (modern machines use algorithms and high powered mathematics to overcome these shortcomings, but still there is only one remotely located sensor)

So right away, at home, particularly if using a nasal mask, we've taken away the ability to directly measure and perhaps accurate depict the Flow Rate Wave - The only thing, really, being sensed. So, we have to be careful when we say, "this looks like that."

Some of your breathing looks like it could include some mouth breathing. The top (early) example is relatively normal, the second (later) example has the characteristically flatter expiratory curve of exhaling through the mouth.

Have a look at the expiratory mouth breathing example in the graphic immediately below. This graphic is important because it uses a only a nasal sensor so the waves are going to be similar to those using a nasal mask. Also, review that other link I sent which uses the other graphic below to explain the effects of mouth breathing on flow. Also, consider the algorithms your machine uses to apply EPAP and IPAP (I don't have a clue and I don't feel like researching it) All of these things will have an influence on the shape of the flow wave graphic.

One thing that you might want to consider - using a FFM with a fixed pressure. That would eliminate some of the unknowns for analytical purposes.


Below: How the shape tends to change with an increase in mouth expiration


And More: Notice at the top right, total exhalation by mouth, the relatively small leak detected. Only about 8

Flow wave examples of mouth breathing wearing a nasal mask:
23:44:00 to 23:44:44 Total mouth breathing on inhale and exhale, 44 second apnea scored at resumption of nasal breathing, no leaks scored.
23:44:50 to 23:45:20 Nasal inhalation with exhalation totally by mouth scored as leak.

---------------------------------------------------------------------------
Below:
Normal sleep breathing on top, expiratory mouth breathing on the bottom in large leak (shaded area)
Notice the flattening of the expiratory portion of the wave with exhalation by mouth.

Edit: Times corrected

[rkl122]

..............

Ron... Chill!

There's nothing to be scared about. You're letting your inquiring mind and imagination run away with you.

You have poor sleep. Lot's of causes for that. Sleep Disordered Breathing is one, caffeine is another, I'm just sayin'. Obsessing over it is a major cause - again, just sayin'.

Thank you Jay! Hey, you may be bringing me good luck. Last night I had my best AHI in 5 months (2.20). Ok, I'm chillin' (but wait'll you see the monstrosities of flow traces I'll link below...)

Yes, I like average better than median most of the time. Statisticians like median, I guess, because it is not affected so much by any large outliers in the sample. ..........

My thinking on that was, for ventilation data, at, say, 15 breaths/min x 60 min/hr x 6 hours/session = 5400 breaths/session, the median is not going to be far from the average. So far, that seems to be the case for me. Whatever. Not my focus now.

Anyway, back to you. You have some Flow Waves that are not "typical". So, we need to investigate why that might be. The first thing is that disturbed sleep from any cause and being awake will produce waveforms that are not typical. The second thing is that we are working with inadequate and incomplete sensing capability with marginal analyzing ability. When we see something in our data, sometimes the best we can do is say, "Oh, that looks, like it might be this". The this being something that came from a lab with a full array of sensors, analytical computers, and experts who have studied the subject for years.

Sleep Labs will likely have at least 12 channels of information to analyze. We have 1, maybe 3, if we have a pulse-Ox..........

Yup, understood. Our flow rate data should be very precise, though, since the sampling rate is so high. The Phillips Resperonics machines are sampling at 125 Hz. (I have a reference somewhere.) BTW, I do have a recording pulse oximeter, and last night it gave an average of 93 with a 95% value of 94. I need to get it calibrated, but the reading accords with the "mild hypoxia" finding in my (sleep lab) sleep study. The doc didn't say a word about it. You can bet this is tops on my agenda for next appt.

............So right away, at home, particularly if using a nasal mask, we've taken away the ability to directly measure and perhaps accurate depict the Flow Rate Wave - The only thing, really, being sensed. So, we have to be careful when we say, "this looks like that."

Some of your breathing looks like it could include some mouth breathing. The top (early) example is relatively normal, the second (later) example has the characteristically flatter expiratory curve of exhaling through the mouth.

Have a look at the expiratory mouth breathing example in the graphic immediately below. .....

Thank you for taking the trouble to post the graphics and refs. I will study them. I do have some mouth breathing, as evidenced by occasional waking, finding myself supine with dry mouth. My significant spooner starts me out secured into side sleeping position, but after she falls asleep, she rolls away, leaving me with a desert of bed to roll onto. Gotta work on that technique...

One thing that you might want to consider - using a FFM with a fixed pressure. That would eliminate some of the unknowns for analytical purposes..........

Yup, I'll start on that tonight. I have an Amara ffm (it doesn't say "View" on it, just "Amara", so I'm not sure exactly which one it is). I'll also turn off the AFLEX. Will bump this thread when I have a few days' data.

So now to some scary(?) pictures. The first link below is the same one in my previous post (for convenience). The second contains first the full same night from which the first link images were taken (oops,sorry it's the third one; didn't rearrange the way I wanted), then four zoomed windows showing what I hope(!) you'll agree is "junkola". I can't be certain I was asleep when they occurred, but I think I was. Note that one of the images shows an apparently 15 sec inspiration without expiration, and wasn't even flagged by the machine. Note also, that the events contain both inspiration and expiration and that there are events of shorter than 10 seconds duration. I get that every night. Is that not a violation of the central algorithmic dogma: 10 sec min, and no flow during the event!? As everyone always says: the machine doesn't know whether I am awake or asleep. But let's leave that for another thread. Observations like this lead me to wonder if my machine is broken. 'Cause if it's not, maybe I am.

http://imgur.com/a/ONWio
http://imgur.com/a/h6iqD

Thanks so much Jay, and anyone else who cares to offer interpretation.

-Ron

[Jay Aitchsee]

Hi Ron,
I've only had time to just glance at your latest postings. Yeah, there's something going on, but I'm guessing some of that wicked stuff is wake and some is mouth breathing. Average, median, it doesn't matter much, here. I do want to study your posts more and I will, but I wanted to comment on a couple of things first. I think I understand your question about less than 10 seconds being an event. But here's the deal, the manufactures have set their own definitions about what is flagged and what isn't and they haven't always agreed. If you have the clinician manual for that machine, check it and see what it says
I also want to direct your attention to that last group of charts above. The first example starts with some "normal" breathing and then flat lines and then a section of mouth expiration. During the flat line I was breathing through my mouth, but the machine didn't recognize it. It did score an apnea at the end, but under some circumstances, it would not have. I did wake once to find the Dreamwear under my chin, the pillow was being blocked by my skin so there wasn't a huge leak. When I woke, I just put the pillow back under my nose and went back to sleep.When I reviewed the data, I saw the machine went to flat line, but didn't score an apnea. There was a leak where I moved the mask back under my nose and that was that. Had I not remembered what had happened, looking at the data, I would have thought I stopped breathing for a few minutes. The point being, you can't always be sure what's happening with our limited number of sensors. The flow rate data, as you point out, probably is precise for the flow passing the sensor at the machine. But after that, we really can't be sure where the "flow" goes.
I think it will be interesting to see what happens with a FFM.