How accurate is the Airsense AHI number?

[drfaust]

So, I'm about 5-6 weeks in to my CPAP therapy and I've been having rough goes at it the last few nights. Rough sleep, don't feel rested, etc. MyAir from Resmed says my AHI hovers between .5 and 3, and that I'm doing great.

BUT...

I get the data off of the SD card and look at it in Sleepyhead and I have two observations:

1) The average for the whole night might be 1 or 2, but there are periods in the night (typically toward morning) when it goes up to 7-9. The number of events increases toward morning.

2) When you scroll through the data looking at small windows of time, there are LOTS of instances where I stop breathing or slow down considerably or the volume drops considerably, but the machine doesn't flag it as an event.

Is this typical? Should I be concerned that my AHI goes high for periods of time, even if the AHI for the whole night is low?

Here are links to some Sleepyhead screen shots that show flagged and non-flagged events. I see lots of similarities between the two:

http://imgur.com/a/fDChQ

I have a follow-up with the sleep doc this week, so I'd love to know if these are issues I should be concerned about.

[Pugsy]

The events you spotted without a flag are most like less than 10 seconds of duration. They have to be at least 10 seconds duration to get a flag.
They also have to have at least a minimum amount of flow reduction.
The one you posted...I don't think it was quite 10 seconds long.

If you are seeing clusters of events at times and not much at other times then your pressure is probably not optimal during those times where you see the clusters.
The 2 most common culprits for needing more pressure sporadically like that are supine sleeping or REM stage sleep.

If the clusters are more in those wee hours of the morning it's possible they are REM stage sleep related because that's when we usually have more REM.

Can you post one of those detailed reports so we can see what you are seeing in terms of those clusters?

[drfaust]

Thanks for the reply.

Yeah -- I figured they didn't meet the 10sec threshold to be flagged, but I'm clearly stopping breathing for chunks of time. If it's 8sec or 10sec, it seems like it's still a problem.

And attached is a screen shot showing clusters. One at the start of the night, and one at the end of the night.

http://imgur.com/TkTvdmA

The email I got from the sleep doc when I inquired about some other thing said that as far as he is concerned, my OSA is "cured" because my AHI is low. He just looks at the bottom line number, which is generally low. The number for my night that night was 2.8. They have my pressure set to be fixed at 8, even though I have an AutoSet machine (the DME gave it to me with the menus disabled, so that they can remotely set it -- they said they basically never set it to auto adjust, and always use the doctor's prescription for constant pressure).

Here is a shot of the AHI change over the night, but the average that gets reported was 2.8

http://imgur.com/uUStHON

[Pugsy]

Can you get me the entire daily detailed instead of a snippet?

Do you ever sleep on your back?
That first cluster is a little early in the sleep cycle for REM.

I want to see the AHI stuff and statistics on the left side and these graphs on the right side
Events
Flow rate
Pressure
Leak

So that it looks like this one

[drfaust]

Here you go:

http://imgur.com/UBKL8RF

I usually fall asleep on my back, but roll to my side in the middle of the night -- I get low back pain and have to pee usually. Rolling on my side helps a bit (I can wait longer before getting up to pee, and back hurts less, but I can't fall asleep on my side).

Edit:

Here's one with the calendar removed, so you can see more info:

[Pugsy]

A little over half of that AHI is central and I suspect that a good chunk of them are awake/semi awake breathing irregulaties getting flagged when you wake up an turn over.
We can't fix centrals/clear airway with more pressure.
You have to mentally remove the centrals/CAs from the AHI evaluation and when you do that your AHI is quite low and the clustering isn't nearly so bad.
Unless the central is a post arousal central because a hypnea or OA or flow reduction causes the arousal and that's when you get awake/semi awake breathing irregularities that can confuse the machine.

How many times do you think you wake up? When you do wake up approx how long to get back to sleep?

Your pressure is 8 cm fixed cpap. If you are waking often...maybe it might be worth considering just a little more pressure to see if reducing the hyponeas and OAs helps with the awakenings and cross your fingers maybe with the centrals if they are related to the arousals.
Your flow limitation numbers aren't very high so I doubt the flow limitation graph is all that busy looking.

The centrals you have to pretty much ignore. We can't fix them with more pressure with your machine (unless they are related to arousals and if we reduce the arousal then they might reduce) and you really aren't having a huge number of any kind of events.

[drfaust]

Thanks for the feedback. I tend to have a lot of nightmares, especially toward morning. So I often will arouse briefly at the end of the nightmare, but fall right back to sleep. I'm not a particularly active sleeper. I fall asleep on my back, but then roll to my side and stay put.

During my titration sleep study the doctor noted "several isolated central apneas." But also noted that a pressure of 8 kept the AHI under threshold. So I think the central apneas might be real, since they were also noted in the PSG in the lab.

But I'm still wondering about my questions:

When I scroll through the data I gave tons of events that seem to be just sub-threshold for the machine to flag them. So they don't contribute to my AHI, but they might be impacting my sleep. And I have periods when my AHI goes high, even though the nightly average is low.

I'm asking because of the nightmare issue. One of the main reasons I was referred to the sleep clinic is because of my constant nightmares. Sleep Doc thought they might be caused by apnea (even though I don't snore -- confirmed by my partner, the PSG, and the machine). Turns out I do have OSA, but the nightmares are still there even with treatment. But when I look at the data, my AHI goes up for periods of time, and I have tons of sub-threshold events that don't contribute to the AHI. These happen at around the same times I have nightmares, which makes me wonder if the "bottom line" AHI number reported by the machine (and which my doc looked at during our phone convo last week when he said the treatment was effective) might be misleading.

[Pugsy]

The medical community had to come up with some sort of line in the sand for the criteria that make up apnea events of some sort.
They came up with the 10 seconds...and yeah there's not a whole lot of difference between 8 or 9 seconds and 10 seconds but it is the way it is.

There is a way in SleepyHead to include those less than 10 seconds flow reductions if you want to see them.
Go to Preferences/CPAP tab and look in the lower left for user customization choices.

Now do these less than 10 seconds events cause a problem? Maybe...maybe not. A lot might depend on just how many you have close together.
Something for you to discuss with your doctor.
If you are having a lot of them...the treatment is a little more pressure to better prevent them from happening.
You might also consider auto mode...if the flow reduction happens the machine might try to prevent the flow reduction with an increase in pressure even with a less than 10 second duration event.

Centrals...an occasional central even if it is the real deal is normal. It's normal to have what we call a sleep onset central and if a person wakes often during the night then that obviously means more sleep onset time and a chance for sleep onset central to pop up. Not a big deal unless we have so many sleep onset centrals we can't really ever make the transition to sleep. Even if every one of your Centrals were the real deal...not enough of them to warrant changing therapy to deal with centrals.

The nightmares...may or may not be related to sleep apnea and the machine only fixes problems related to sleep apnea.
As much as we want to put everything that bugs us in the sleep apnea basket and have cpap fix it....cpap can't fix problems not related to sleep apnea and there's a long list of sleep problems out there that have nothing to do with sleep apnea.

If your nightmares are related to sleep apnea and in your case a possible not quite optimal therapy due to the number of flow reductions that don't meet the 10 second rule...then it sure wouldn't hurt to try to reduce those flow reductions with more pressure. It won't hurt to at least try. Will it help? Dunno but it's worth at least trying.

[Biguugly]

FWIW, been having nightmares for 40 years. All night, every night. Been on cpap for 2years, and have not seen any difference. MD and I found out by sheer accident that mild sedative (clonazepam 2mg) reduced the vividness of nightmares considerably. Still have very unpleasant dreams, but not heart pounding, jolt awake nightmares. Hope you can find help. This sucks!

[drfaust]

Thanks all -- this feedback is fantastic. I'll try adjusting the thresholds in Sleepyhead and see what happens.

And yes -- I've found that .5-1mg of clonazepam does wonders for the nightmares. I used to wake up panicking constantly and screaming (my partner obviously did not love that) multiple times per night. With klonopin I sometimes have mildly stressful dreams, but never to the level that I had previously. I just don't want to become dependent on/tolerant to benzos over the long term. I'm hoping for a non-pharma fix eventually, but I'll take the drugs in the meantime if they mean not waking up screaming.

The sleep doc thought that maybe the nightmares were triggered by apneas, since apparently that's quite common. It seems they aren't. But I'm still happy I'm getting therapy that will help me long-term with cardio issues, blood sugar regulation, etc. I'm not a fan of the hose, but I'm a fan of fixing all of the other problems that the hose fixes.