Calculating non-Supine Sleep Apnea Metrics

[bv1800]

I finally requested a copy of my sleep study results. I've always been a side sleeper (at my wife's insistence), because I snore excessively when I sleep on my back (Supine). In my study results, my overall RDI is 60.2 (AHI 57.2), so clearly, I had severe sleep apnea. The report states that I spent 84.3% of the sleep time in the supine position (apparently, non-supine time was right side only). My RDI in the supine position was 71.7. If I assume that the overall is the average for the full study, I can mathematically calculate the non-supine value via the following formula:

Overall RDI = Supine RDI X Supine Time Pct + non-Supine RDI X non-Supine Time Pct

Plugging in the numbers:

60.2 = 71.7 X 0.843 + non-Supine RDI X (1-0.843)

So

non-Supine RDI = -0.36 (or essentially zero with the numeric values being caused by rounding errors). These calculations look ridiculously low until you realize that 71.7 X 0.843 = 60.44, which would mean that all of the events measured occurred in the supine position.

If this calculation is valid, then it would appear that I have OSA in the supine position only. Does anyone have any idea if my calculations are valid? From researching the topic, I've found some sample sleep study results, where overall, supine and non-supine data are available. The calculations works as I've performed them, then I think a new conversation with my Dr is needed.

[Julie]

What you need is to try sleeping only on your side for a week, and then looking at your software stats to see where your AHI was. What you've done is nice math, but it's theoretical, not hands-on proof. You may be absolutely right that you have only positional apnea, but you need to give it a good solid real life try before deciding to change things on your own.

[bv1800]

What you need is to try sleeping only on your side for a week, and then looking at your software stats to see where your AHI was. What you've done is nice math, but it's theoretical, not hands-on proof. You may be absolutely right that you have only positional apnea, but you need to give it a good solid real life try before deciding to change things on your own.

... and there's the rub....

I was issued a brick. I didn't find this site until it was too late to switch. I meet with my Dr in Sept. At that point in time, I'll bring this evidence to him. I'm just attempting to determine whether or not the way that I arrived at my "theoretical" number is a valid way to estimate the value. I won't change a thing until I get proof from actually testing the hypothesis.

Thanks for the response.

[Pugsy]

The only real confirmation to your theory would need to come from a sleep study where you slept on your side entirely.
If you had a full data machine you might get an idea of what is happening if you set the machine to the minimum pressure of 4 and slept on your side entirely but then that is just a maybe because even 4 cm pressure has therapeutic value so we wouldn't know for sure if you didn't have any events because you simply didn't or you didn't because the minimum pressure dealt with them effectively.

[JDS74]

The number of events in each sleeping position is often reported in the sleep study.
Do you see anything there that gives you the total number of events while in non-supine position?

[bv1800]

The only real confirmation to your theory would need to come from a sleep study where you slept on your side entirely.
If you had a full data machine you might get an idea of what is happening if you set the machine to the minimum pressure of 4 and slept on your side entirely but then that is just a maybe because even 4 cm pressure has therapeutic value so we wouldn't know for sure if you didn't have any events because you simply didn't or you didn't because the minimum pressure dealt with them effectively.

An excellent point. At this point, even results from 4 CM would be good to have, since I would expect very low numbers of events, if my hypothesis/calculations are true (and definitely, refute them, if 4CM didn't resolve my events). I won't stop therapy until I have a valid study.

The number of events in each sleeping position is often reported in the sleep study.
Do you see anything there that gives you the total number of events while in non-supine position?

Sadly, no. It only stated RDI overall and in the supine position. This is the reason that I've estimated the value via the calculation that I outline.

[Pugsy]

You might ask your DME if they had a full data Respironics or ResMed machine that they would loan/rent you to try so you could see what might be happening at 4 cm on your side (assuming you can for sure stay on your side).
You would only need a few nights (maybe just one) to get a maybe idea.
One night if you ended up with a lot more apnea events happening because the pressure was insufficient.
Do you know anyone that you could borrow a full data machine from?

Another thought ....sometimes our OSA is worse in REM sleep and not so exciting in non REM sleep (that's me) and if you were one of those people and didn't happen to be in REM sleep when you were on your side during the sleep study that might impact the potential AHI.

[bv1800]

Another thought ....sometimes our OSA is worse in REM sleep and not so exciting in non REM sleep (that's me) and if you were one of those people and didn't happen to be in REM sleep when you were on your side during the sleep study that might impact the potential AHI.

Great minds think alike I sent a note to my DME, requesting one, a few minutes ago.

I recently lost more than 10% of my body weight and was suffering severe mouth breathing (painful aerophagia when I taped my mouth shut for a one night test). My contact told me that they had auto-titrating machines that they could loan for 2 weeks, if I wanted to do the study, instead of trial and error adjustments. I'm thinking that it's time to do so.

[Julie]

The first place most people lose weight from (so that it's obvious) is their face, so you may need a new mask, or a real refit of your own.

[bv1800]

The first place most people lose weight from (so that it's obvious) is their face, so you may need a new mask, or a real refit of your own.

I use the AirFit P10. No need for either.

On another note....

I was going to leave this alone since I had the info that I was seeking, but since I'm writing a response, I want to include the following.

I can understand people's attitudes that the calculations that I did are "theoretical", but they in fact are as precise as any other discussion in this forum. To whit:

Total events = # events supine + # events non-supine
RDI = Total Events/Sleep Hours = # events supine/Sleep Hours + # events non-supine/Sleep Hours
= (# events supine/supine sleep hrs)X(supine sleep hrs/Sleep Hours) + (# events non-supine/non-supine sleep hrs)X(non-supine sleep hrs/Sleep Hours)
= Supine RDI X Supine Pct + non-Supine RDI X non-Supine Pct

so

RDI = Supine RDI X Supine Pct + non-Supine RDI X non-Supine Pct

This isn't a "math trick". It's only a "theoretical" as RDI itself.

Said differently:

If my overall RDI is 60.2 and I slept for 6 hours during the study, then the number of total events is 6*60.2 = 361
if my supine RDI is 71.7 and 84.3% of the time I was in the supine position, then my time in supine position is .843*6 = 5.058 hrs, so the number of events in supine position is 71.7*5.058 = 362.

While the numbers don't match exactly, due to rounding of the reported metrics. This is precisely how the machines calculate the scores, so this derivation is a mathematical proof of what my non-supine numbers are, assuming that the RDI scores are for the full reported times (overall and while supine, and not some portion of the reported times).

I apologize for going all "nerdy" on the readers, but I do this "math stuff" for a living.

[palerider]

While the numbers don't match exactly, due to rounding of the reported metrics. This is precisely how the machines calculate the scores, so this derivation is a mathematical proof of what my non-supine numbers are, assuming that the RDI scores are for the full reported times (overall and while supine, and not some portion of the reported times).

I apologize for going all "nerdy" on the readers, but I do this "math stuff" for a living.

the only problem I see with that is the same problem with trying to wrap up a whole night in one tidy number...

the rates of rdi in different positions during the night are quite likely to be different, much higher supine vs non.

but, like you said, probably as good as the other 'summation' numbers.

be interesting to see the standard deviation

[bv1800]

the only problem I see with that is the same problem with trying to wrap up a whole night in one tidy number...

the rates of rdi in different positions during the night are quite likely to be different, much higher supine vs non.

but, like you said, probably as good as the other 'summation' numbers.

be interesting to see the standard deviation

I think that it also begs for multi-night testing and very explicit testing both supine and non-supine as well as REM and non-REM, including all 4 pairings that come from matching up each of the two (what statisticians would call a "full experimental design"). I'd wager that the lack of a full experimental design in testing is the reason that the vets on this forum spend so much time helping people get things dialed in. I point to mikesleep who's new to CPAP therapy and getting great numbers, from using pressures that are half of what was prescribed and prescribed pressures are untenable. I wouldn't doubt that his treatment pressure was estimated while sleeping supine (who can sleep on their side with a metal bolt attached to their temple), but he's a side sleeper, so in actual practice, his pressure need is vastly lower than what was prescribed.