UARS

[Sleepster]

OSCAR provides information about the AHI, which is the sum of the OA, CA, and hypopnea indices. But does it provide any information about the RDI, which is the sum of the AHI and the UARS index? Or, once you're on PAP therapy and your AHI is under control, you don't need to concern yourself with UARS because if the PAP therapy is successfully treating OA's and hypopneas, it must also be treating any UARS that might have been present before PAP therapy was started?

[Crimson Nape]

From my rough understanding, only the Autoset provides this information. Since you are using the VAuto, I don't think it has that capability.

- Red

[Sleepster]

Thanks. I can see how UARS would cause problems pre-PAP therapy, and those problems would be compounded by a sleep study or diagnosis that that didn't take UARS or the RDI into account. But I would think that with PAP therapy that successfully treats OA's and hypopneas, UARS would not be a concern.

I also wonder how PAP machines, like the Autoset, can measure UARS.

Perhaps this thread should be moved to the main forum where more people might see it.

[OpalRose]

Perhaps this thread should be moved to the main forum where more people might see it.

Thread has been moved to the Main Forum.

[unadvisedfun]

In UARS patients RERAs are usually the problem. The aircurve 10 Vauto and S do not flag RERAs.
airsense 11 autoset isn’t much better… it claims to flag RERA’s but in practice would only catch about 5%-10% of my actual RERA’s. It may flag 1 or 2 a night but I was having about 50+ a night. 

I was thinking about contributing a user-defined RERA flag to oscar. User would enable similar to the existing UF1 and UF2 flag. Criteria would look for a sudden increase in inspiratory flow rate (like a big gasp) or pattern of faster & deeper recovery breathing.

This would allow UARS patients to get a better view of their subtle RERAs and and give visibility to folks who have a machine that either doesn't flag RERAs or misses RERAs.

[McPhearson99]

When you say "sudden increase in flow rate" do you mean a spike like the one in the attached screenshot? I'm asking because I have a lot of these every night but my AirSense 11 doesn't flag them as anything.

[Sleeprider]

Even the machines that flag RERA are probably under-counting respiratory event related arousals because they have no means of measuring sleep stage and arousal. The surrogate is an increasing flow limitation ending in recovery breathing which is inferred as arousal. I find the machines poorly measure flow limitation when the baseline respiration is flow limited, but it does pretty well when breathing varies between unobstructed normal respiration and inspiratory flow limitation. The persistence of flow limitation in UARS patients means they are the ones that have a lower flow limitation index in spite of more persistent flattened inspiration. I don't rely on the machine to report this metric accurately unless verified through evaluation of zoomed images of the flow rate. If flow limitation never changes, even severe flow limitation does not register on Resmed's summary data and RDI will vary among individual, sometimes with serious under-counting being prevalent in those with persistent chronic flow limitation.

That's my working theory on the matter. I'd like to hear the observations of other members.

[unadvisedfun]

When you say "sudden increase in flow rate" do you mean a spike like the one in the attached screenshot? I'm asking because I have a lot of these every night but my AirSense 11 doesn't flag them as anything.

Yes, it could be. Would need to have EEG (brain wave monitoring) to know for sure.

[McPhearson99]

The persistence of flow limitation in UARS patients means they are the ones that have a lower flow limitation index in spite of more persistent flattened inspiration.  I don't rely on the machine to report this metric accurately unless verified through evaluation of zoomed images of the flow rate.

I agree. In my experience, I've noticed that in the minutes leading to a wakeup with a gasp and jolt of adrenaline the shape of the peaks in my flow rate graph get noticeably flatter and start showing double peaks. I don't think my AirSense has flagged a single RERA in the year or so I've been using it.

Yes, it could be. Would need to have EEG (brain wave monitoring) to know for sure.

I don't know of any any home EEG devices unfortunately but if I ever go for another sleep study I'll try to get the raw data for it.
In the meantime, I'm trying to use Python and SciPy to find outliers in the peaks of my flow rate data. So far I've noticed that they seem to correlate with REM sleep, which I expect in my case. I used EDFbrowser to export the device data to CSV but there's libraries available to read EDF files in Python.
I've also noticed that a double peak (peaks less than 1 second apart) where the second peak is higher than the first might be a good indicator of an RERA. I've attached a few screenshots of my flow graphs with the double peaks marked with Xs. The X axis is the index of each datum and the Y axis is the flow rate in liters per second (1/60th of OSCAR's Y axis).

[Muggzy]

I think SleepRider is exactly right regarding the criteria that the machine uses to identify an arousal in the absence of EEG data, where they use a "recovery breath" as a surrogate to detect an arousal. In fact, I feel like I've read exactly that in one of the ResMed papers, but cannot find it now.

Here's a visual depiction of my understanding of it, anyways: https://youtu.be/ftyrSjcMqGc

There are several papers out there describing various mathematical methods of detecting flow limitation that look like they might be more robust than ResMed's and work even in the case where flow limitation is the norm rather than the exception, but most of them are well above my pay grade.

I do wonder if a simplified model using peak finding and average slope determination might suffice though, since some of the most common visual features of flow limitation include one or more peaks whose width is significantly shorter than the single peak of an NIFL breath should be, and an average slope between the 20% and 80% points of the inspiration that doesn't change direction by nearly as much as a normal inspiratory peak would.