10 sec apneas on ResScan?

[rada]

I feel comfortable to conclude after a full month of monitoring with ResScan software, oximeter, and webcam that a moderate OSA based on AHI is not necessarily coincident with reduced SpO2. For instance, my ResScan event map from this morning shows a AHI of 15.8 but the percent time in apnea was only 1% and my SpO2 averaged 97.8%. Most of my apneic events were 10 seconds long. My point is that AHI alone is not an adequate measure of sleep difficulty. I have bad nights with AHI of 15 and low SO2 but not last night.

[rada]

I am feeling that maybe the therapy is working for me. I have struggled for years with this. These are last two nights which also highlight the frequency of 10-second apneas -- a 10-sec phenomenon not yet seemingly well explained.
Here are apnea events from two nights ago Feb 25:


and the associated SpO2:


Here is the ResScan data from Feb 26:


with the associated SpO2:


I see only a very slight negative impact, if any (maybe the negative impact is from the hyponeas separate from the apneas), of 10-sec apnea on SpO2. If anyone wants to see other charts showing sleep position, pressure, and such then just ask.

[roster]

I suspect something is wrong in the program. If one sees that many 10-sec apneas, would you not also expect to see an occasional 11- or 12-second apnea?

[jnk]

I suspect something is wrong in the program. If one sees that many 10-sec apneas, would you not also expect to see an occasional 11- or 12-second apnea?

All I can figure is that some sort of pattern involving cycling episodes of shallow breathing is occurring, but the decreases in ventilation are occurring right on the edge of the 75% line that ResMed uses as the dividing line between hypopneas and apneas, so that occasionally, but regularly, one of those hypopneas is crossing the line and getting scored as an apnea of the minimum duration. They have got to be central in nature, in my uneducated opinion. I can't see how in the world obstructive events could be anywhere near that consistent in that way.

But, hey, what do I know. Maybe, like you say, there is some kind of bug.

jeff

[rada]

Dear Roster and Jeff,
I appreciate your feedback. I agree that the explanations might relate to
• the algorithm,
• my physiology, or
• some combination of the preceding.
Roster suggests my program may be flawed. My machine seems to work reliably in most cases that I can judge; for instance, when I set the settling time, the machine settles for that amount of time. Jeff suggests that the normal working of the algorithm might have crossed into some oddity of my physiology so as to produce such a record. He notes that the hyponea algorithm considers a 75% reduction in flow to be a hyponea. I am reading now the clinical manual for the machine. It seems to say on page 11:
• an apnea is scored: if the 2-second moving average ventilation drops below 25% of the most recent 100-second average for, at least, 10-consecutive seconds.
• a hyponea is scored: if the 8-second moving average drops below 50% but not below 25% of the most recent 100-second average for 10 consecutive seconds.
As the apnea proceeds, the most recent 100-second average will be driven lower by the ongoing apnea and that would seem to bias the result unless most recent means the most recent before the beginning of the reduced airflow. Jeff, might you explain what you were saying about "right on the edge of the 75% line that ResMed uses as the dividing line between hypopneas and apneas, so that occasionally, but regularly, one of those hypopneas is crossing the line and getting scored as an apnea of the minimum duration"? I was trying to imagine something like
• the ventilation drops below 25% and signals to the machine the beginning of an apnea, but
• before the 10-seconds have passed, the ventilation improves enough to move into the hyponea category, but
• the algorithm mistakenly keeps the apnea label, but
• assigns a duration of 10-seconds since that is somehow associated with the duration for hyponeas.
I could imagine an error like this could occur. However, then I would not be the only person to experience this error, right? I get sometimes event graphs that look like other people's graphs without any 10-second apneas.
Jeff, you said: "They have got to be central in nature, in my uneducated opinion. I can't see how in the world obstructive events could be anywhere near that consistent in that way." Why could central apneas be consistent in this way but obstructive apneas not?
Thank you,
Roy

[jnk]

Dear Roster and Jeff,
I appreciate your feedback. I agree that the explanations might relate to
• the algorithm,
• my physiology, or
• some combination of the preceding.
Roster suggests my program may be flawed. My machine seems to work reliably in most cases that I can judge; for instance, when I set the settling time, the machine settles for that amount of time. Jeff suggests that the normal working of the algorithm might have crossed into some oddity of my physiology so as to produce such a record. He notes that the hyponea algorithm considers a 75% reduction in flow to be a hyponea. I am reading now the clinical manual for the machine. It seems to say on page 11:
• an apnea is scored: if the 2-second moving average ventilation drops below 25% of the most recent 100-second average for, at least, 10-consecutive seconds.
• a hyponea is scored: if the 8-second moving average drops below 50% but not below 25% of the most recent 100-second average for 10 consecutive seconds.
As the apnea proceeds, the most recent 100-second average will be driven lower by the ongoing apnea and that would seem to bias the result unless most recent means the most recent before the beginning of the reduced airflow. Jeff, might you explain what you were saying about "right on the edge of the 75% line that ResMed uses as the dividing line between hypopneas and apneas, so that occasionally, but regularly, one of those hypopneas is crossing the line and getting scored as an apnea of the minimum duration"? I was trying to imagine something like
• the ventilation drops below 25% and signals to the machine the beginning of an apnea, but
• before the 10-seconds have passed, the ventilation improves enough to move into the hyponea category, but
• the algorithm mistakenly keeps the apnea label, but
• assigns a duration of 10-seconds since that is somehow associated with the duration for hyponeas.
I could imagine an error like this could occur. However, then I would not be the only person to experience this error, right? I get sometimes event graphs that look like other people's graphs without any 10-second apneas.
Jeff, you said: "They have got to be central in nature, in my uneducated opinion. I can't see how in the world obstructive events could be anywhere near that consistent in that way." Why could central apneas be consistent in this way but obstructive apneas not?
Thank you,
Roy

Hi Roy,

As I understand it, mathematically speaking, a drop below 25% of the recent average would be a drop of greater than 75%. I think that is why the clinical guide in the sentence before the one you quoted says in describing its scoring techniques:

"An apnea is defined as a greater than 75% decrease in ventilation."

That is the line, since for hypopneas, the clinical guide says:

"A hypopnea is defined as a 50 to 75% drop in ventilation."

That is the sole basis for my hypothesis, based on nothing more than my uneducated assumptions and partial knowledge, that you may be having irregular breathing in which the change in ventilation falls between 50 and 75% of the sliding average most of the time and is only occasionally, relatively speaking, crossing the 75% line. That might make the numbers and charts look odd as far as so many of the "apneas" being exactly 10 seconds in length, when in actuallity, it is a matter of highly variable "events" occasionally crossing a line in such a way as to be scored as a 10-second apnea in the machine's estimates of events.

Some central-type events can have a regular cycle to them, underbreathing following overbreathing following underbreathing over and over again, whereas purely obstructive events are not known for having that kind of cycling regularity. That is what I meant by my assumption that some kind of cycle of overbreathing and underbreathing must be showing up as those odd 10-second apneas over and over throughout much of the night. I am not assuming you have some form of central sleep apnea. I am just assuming that it must not be truly, solely, obstructive. And I admit I'm not really in a position to make that kind of assumption. I just think that is more likely than that the machine is defective in some way.

I am not a professional and I am relatively new at this. It would take someone like -SWS, who understands the algorithms MUCH more completely than I do, to give you any input with any real scientific basis. I was only making a guess based on assumptions. Sometimes I'm right by accident. But often I am wrong. So file it accordingly.

jeff

[rada]

Hi JNK,
I understand that a drop below 25% of the recent average is a drop of 75% (if the price of Valentine candy at the store today is 25% of the Feb. 21st price, then the price has dropped 75%). You said:
"that you may be having irregular breathing in which the change in ventilation falls between 50 and 75% of the sliding average most of the time and is only occasionally, relatively speaking, crossing the 75% line. That might make the numbers and charts look odd as far as so many of the "apneas" being exactly 10 seconds in length, when in actuality, it is a matter of highly variable "events" occasionally crossing a line in such a way as to be scored as a 10-second apnea in the machine's estimates of events".
If my breathing is often moving into the realm of a 50 to 100% drop and hovering around the 75% drop line often --- ahah! --- now I think I see your point. Are you suggesting that if my breathing were somehow hovering around '75% drops' much of the time, then they might usually be in the hyponea category but just occasionally stay below 75% long enough to make it to 10 seconds but seldom have the wherewithal to make it to 11 seconds? I am trying to imagine the model of a time series that would do that:
• It drops below 75% for a second before rising above 75%.
• Later it drops below 75% for 2 seconds before rising above.
• It would have to stay below 75% for 10 seconds to mark it an apnea.
If we imagine the probability of the 75% drop being 0.5 and the events being independent of one another, then the probability of a 2-second drop would be 0.5 * 0.5 = 0.25. If that model were accurate, then I should have
• half as many 11 second apneas as 10 sec,
• half as many 12 sec as 11 sec, and
• so on.
The probability of a n-second drop would be 0.5**n. Would the preceding charts for Feb 26 and 25 fit such a model?
I see your point about different physiological models for the low-duration apneas than the high-duration ones. If all the apneas were a function of this hovering around 75% drop phenomenon, then the odds of one 45-second apnea, as shown on Feb 26, but the next longest apnea being 30 seconds seems slim. I know next to nothing about central apneas. I did not realize that they tend to a greater periodicity or constancy than obstructive ones. Without knowing, I had been conjecturing the opposite from what you conjectured. I was conjecturing that the long-duration apneas might be central phenomena and the obstructive apneas were the short duration ones. My naive rationale went as follows:
My reacting to the respirator might occasionally distort my CO2 enough to lead the brain to shut-down breathing for close to a minute (i.e., central). My radiation neuropathy causes fasciculations in my muscles secondary to fibrosis of the nerves. I see these fasciculations in my trapezius and sternocleidomastoid. I have not been monitoring nor do I know how to monitor fasciculations inside my throat or genioglossus (lower tongue). Maybe frequent fasciculations of the muscles of the upper airway were contributing to the phenomenon of drops hovering around 75%, as you described. I wish I had more data and better models and could understand and manage my OSA better. Thank you for enlightening me about the 'hovering' and about central apneas and generally for raising some interesting possibilities.
Yours,
Roy

[jnk]

My main point is that it could be (with the stress on "could" more than "be") that there are really events occurring with a lot of variation, but it only looks like a bunch of them are all happening exactly the same because of the cutoff line of the way the machine scores. Some events are 50%, others 60%, and others 75% and everywhere in between. But since no data is given on that, all that shows up is an estimated hypopnea. My thought is that some of your hypopneas are simply crossing the line into apnea territory at 76%, or something, according to how the machine scores them. So there may appear to be a pattern where one would not be seen in a real PSG sleep study that can truly score the apneas and hypopneas for what they are. But in my opinion only a lab sleep study would show what was really going on because of the limitations of machine-estimated scoring.

I don't think your way of looking at it, or rooster's, is any less probable than mine. So don't put too much stock in my mental wanderings.

jeff

[roster]

I wish I had more data and better models and could understand and manage my OSA better. Thank you for enlightening me about the 'hovering' and about central apneas and generally for raising some interesting possibilities.
Yours,
Roy

No matter how much data we get, we still have a shitty anatomy for breathing while asleep and a very primitive therapy to help deal with it.

You may be overthinking it. But if that is what you enjoy, carry on and keep posting.

[rada]

Perhaps I am over-analyzing the situation in some ways, but the personal, practical repercussions have been substantial. Some of the things that I have learned over the past month of experimenting with position, pillows, chin straps, cable heating, diet, time in bed, and so on include:
• I need the APAP to keep SpO2 above 95%;
• so long as I use my APAP, supine with one pillow is fine for AHI and SpO2, and I feel much better in the end than when I was forcing myself to be on my side;
• the AHI in the teens that I typically get on APAP LED is not necessarily meaningful in terms of SpO2 or quality of sleep. Particularly misleading can be the impact of apneas that are 10 seconds long which may be a physiological fluke. More importantly than AHI might be % time in apnea.
• I need the chin strap in a certain way to reduce the leaks;
• I have started to detect patterns between pulse, SpO2, motion (monitored via webcam), and my mentation (to the extent I can remember how it was during the night) that intrigue me about adding EEG monitoring with the fantasy of getting some control over the quality of my sleep and training myself to exploit neural plasticity.
I hope these dialogs are helpful not only to me. I believe patients can empower themselves through sharing information. The medical profession does not have some of the resources that we uniquely bring to the table. Thanks for the feedback.
Roy

[KatieW]

Rada, thanks for sharing your "research". I find it interesting, and have learned from it.

[rada]

Katie,
Thank for you the support.
Walking to school after posting my note, I remembered another thing that I have learned about position, but then thought about the overall experience further and wanted to add these two items to my list of an hour ago of what I have learned from monitoring sleep and discussions on this forum:
* The modified (half swastika) prone position is the only prone position in which I can comfortably sleep, but my SpO2 drops by about 2% as compared with the supine position.
* Most of all, monitoring and successfully modeling some of my sleep has led me to realize that I can control some of the outcomes. I once felt that going to bed with APAP was the beginning of a nightmare -- that someone had dropped me involuntarily into a bottomless pit. Now I can imagine at times that each night is an opportunity for a kind of nature hike in sleep land.
Yours,
Roy

[roster]

Now I can imagine at times that each night is an opportunity for a kind of nature hike in sleep land.

Good for you. For me it is more like a stroll through a deep jungle with sharply-fanged predators, stinging and biting insects, and venomous snakes waiting behind every thorny bush. One little mistake, in the form of a mask leak or rolling into a supine position, and the critters get me.

[rada]

Roster, Since your experience is often mine -- a horrible one -- I am meeting in the morning with a prosthodontist to try to get a mandibular advancement device as an adjunct to my APAP and maybe for a chance to get some sleep without choking and without being a complete slave to the APAP. I don't hold much hope for the oral appliance, but to avoid the critters of the APAP-OSA night I would walk a thousand miles through desert. Roy

[roster]

Rada,

I think that is an excellent move. Let us know how the experience goes. There are some members who use both CPAP and MAD. jnk comes to mind.

I am not going to suggest a tracheotomy for you, but I do suggest you read Deadly Sleep, http://www.amazon.com/Deadly-Sleep-Your ... 996&sr=1-1 , in which Dr. Mack Jones describes his excellent experience with his skin-lined tracheotomy.

For myself, I ideate about a situation in which a skin-lined tracheotomy might be appropriate. This would be in case of an injury or health problem that would make it very difficult for me to manage my CPAP therapy. The probability of such a situation arising is higher for me because CPAP is ineffective for me in the supine position.